Skip to content

Content Sweep MCP server

ai.contentsweep/agent-accesscontentsweep.aistreamable-httpv1.0.0Scanned

Prepaid copyright investigation and authorized enforcement, with a free agent sandbox.

35 tools: 4 can delete or cancel things, 2 do something that can't be undone, 18 change things and 11 only read.

Tools: 4 delete, 2 irreversible, 18 write, 11 read.
  • Reads private data · 1
  • Reads text others write · 35
  • Can send data out · 2
  • 10 findings

Tools listed without sign-in. Remote posture

With Anlyon

4 tools here can delete or cancel things and 2 do something that can't be undone. Govern them.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the Content Sweep MCP server can do

The Content Sweep MCP server is published by contentsweep.ai, a domain the MCP Registry verified. It has 35 tools. It can delete or cancel things (4 tools: submit_takedown, revoke_authority, revoke_credential and revoke_webhook), do things that cannot be undone, such as sending, paying or publishing (capture_evidence and purchase_credits) and create or change data (18 tools: find_copies, investigate_url, prepare_takedown, check_status and 14 more). 11 tools only read.

1 tool reads private data (get_account); 35 tools bring in text other people can write (find_copies, investigate_url, capture_evidence and 32 more); 2 tools can send data out of your system (create_quote and create_webhook). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.

The fixed checks matched 10 findings (annotation-conflict and credentials-mention). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in.

Last scanned Nov 11, 2026, version 1.0.0.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

35 tools

Deletes or cancels 4Claude Code: deny
  • revoke_authorityIrreversibly revoke an owned delegation by ID.Reads text others writeNo findings

    Why it is in this group

    • The name "revoke_authority" says "revoke": a word for removing or ending something.
    • The description says "Irreversibly".
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Irreversibly revoke an owned delegation by ID. Requires rights scope. Queued work and future monitor runs recheck revocation; this stops later authorized actions but does not retract a notice already sent. Returns revoked=true on success.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: false
    • openWorldHint: true
  • revoke_credentialImmediately revoke an owned service-principal credential by ID.Reads text others write1 finding

    Why it is in this group

    • The name "revoke_credential" says "revoke": a word for removing or ending something.
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "...ately revoke an owned service-principal credential by ID. Requires credentials scope. Its..."

      What this check looks for
    Description, as the server wrote it
    Immediately revoke an owned service-principal credential by ID. Requires credentials scope. Its dependent access tokens also stop authenticating; revoking the credential used for this call prevents later calls with it. Returns revoked=true.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: false
    • openWorldHint: true
  • revoke_webhookDeactivate an owned webhook and cancel its pending deliveries.Reads text others writeNo findings

    Why it is in this group

    • The name "revoke_webhook" says "revoke": a word for removing or ending something.
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Deactivate an owned webhook and cancel its pending deliveries. Requires webhooks scope; an in-flight delivery may finish. Returns revoked=true. Does not delete case state or events.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: false
    • openWorldHint: true
  • submit_takedownSubmit a prepared packet through a supported, activated enforcement route.Reads text others writeNo findings

    Why it is in this group

    • The name "submit_takedown" says "submit": a word for a change.
    • The server marks it destructive (destructiveHint: true).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Submit a prepared packet through a supported, activated enforcement route. First create_quote(action='submit', input={rights_holder_id, work_id, delegation_id, case_id, packet_id}). Rechecks current delegation, appointment permission, evidence, route, policy and funding before delivery. Results distinguish submission status, provider receipt and removal status. If submission_reconciliation_required is returned, reconcile the original operation before any retry; provider acceptance does not prove removal. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: true
    • openWorldHint: true
Can't be undone 2Claude Code: ask
  • capture_evidenceCapture a new immutable metadata observation for an existing case.Reads text others writeNo findings

    Why it is in this group

    • The name "capture_evidence" says "capture": a word that usually means money moves, which cannot be taken back.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Capture a new immutable metadata observation for an existing case. First create_quote(action='evidence', input={rights_holder_id, work_id, delegation_id, case_id}). Results include evidence_id, SHA-256 hash and timestamped metadata; use get_evidence to export it. Captures no screenshots, images, audio or video. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • purchase_creditsPurchase organization-bound prepaid service credits in USD cents.Reads text others writeNo findings

    Why it is in this group

    • The name "purchase_credits" says "purchase": a word that usually means money moves, which cannot be taken back.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Purchase organization-bound prepaid service credits in USD cents. Requires billing scope and an idempotency key. A live 402 response returns an MPP card-payment challenge; retry the same purchase/key with payment_authorization separately from the account Bearer token. Authenticated provider confirmation credits the balance exactly once. Sandbox purchases are simulated. Cards are supported; stablecoins are not enabled.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Changes things 18Claude Code: ask
  • check_statusFetch a fresh content observation for an existing case.Reads text others write1 finding

    Why it is in this group

    • The name "check_status" says "status": a word for reading.
    • The server marks it as not read-only (readOnlyHint: false).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "check_status" as not read-only (readOnlyHint: false), and its words say it reads.

      Matched text

      annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "check_status" says "status": a word for reading.

      What this check looks for
    Description, as the server wrote it
    Fetch a fresh content observation for an existing case. First create_quote(action='status', input={rights_holder_id, work_id, delegation_id, case_id}). Results include observed status, evidence_id, checked_at and HTTP status. Blocked retrieval or a timeout cannot prove removal. This is paid network work; use free get_operation to poll queued work and get_case to read stored observations. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • create_quoteCreate a free quote binding one action, complete input, rate version and maximum USD-cent price for 15 minutes.Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "create_quote" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Create a free quote binding one action, complete input, rate version and maximum USD-cent price for 15 minutes. Does not reserve or spend credits and does not grant rights. Returns a quote ID for the matching task tool: search/find_copies, investigate/investigate_url, evidence/capture_evidence, prepare/prepare_takedown, submit/submit_takedown, status/check_status or monitor/monitor_case. Authenticate and include the exact rights-holder, work and delegation scope.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • create_rights_holderRegister a rights holder managed by this organization, without a creator login or subscription.Reads text others writeNo findings

    Why it is in this group

    • The name "create_rights_holder" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Register a rights holder managed by this organization, without a creator login or subscription. Requires rights scope. Supply a person/business name and optional country/contact email; returns a rights-holder id for create_work and delegate_authority. Registration records identity and does not itself authorize enforcement.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • create_webhookRegister a public HTTPS receiver for this organization's signed lifecycle events.Reads text others writeCan send data outNo findings

    Why it is in this group

    • The name "create_webhook" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Register a public HTTPS receiver for this organization's signed lifecycle events. Requires webhooks scope; returns the signing secret once. Verify HMAC-SHA256 signatures, timestamp freshness and stable event IDs. Delivery is at least once with bounded retries; private-network destinations and redirects are refused. At most five active receivers per organization.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • create_workRegister a protected original work under an existing rights holder in this organization.Reads text others writeNo findings

    Why it is in this group

    • The name "create_work" says "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Register a protected original work under an existing rights holder in this organization. Requires rights scope. Supply rights_holder_id and description, optionally original_url and your external_id. Returns a work id for delegation and quoted work; registering an original does not determine whether a suspected copy infringes.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • delegate_authorityRecord the rights holder's existing authority for exact works and actions with a reference, attestation version and future expiry.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Record the rights holder's existing authority for exact works and actions with a reference, attestation version and future expiry. Requires rights scope and rights_authorized=true. Returns a delegation id for create_quote. Set can_appoint_content_sweep=true only when the mandate permits appointment; submission requires it. Delegation can be revoked immediately; payment never substitutes for authority.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • enroll_agentCreate a free, isolated sandbox organization without browser signup or authentication.Reads text others writeNo findings

    Why it is in this group

    • The description's first verb is "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Create a free, isolated sandbox organization without browser signup or authentication. Returns organization_id and a bootstrap access_token once; send it as Authorization: Bearer for subsequent private calls. All sandbox credits, search results and submission receipts are simulated; no live providers or real funds are used.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • enroll_liveCreate a separate live organization using an authenticated sandbox account.Reads text others write1 finding

    Why it is in this group

    • The description's first verb is "create": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credentials".

      Matched text

      description: "...ctivation gates apply. Returns new live credentials once. Does not purchase credits or esta..."

      What this check looks for
    Description, as the server wrote it
    Create a separate live organization using an authenticated sandbox account. Requires accountable customer identity, accepted current terms and authorized-agent attestation; supported countries and activation gates apply. Returns new live credentials once. Does not purchase credits or establish copyright authority.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • exchange_tokenExchange an authenticated bootstrap credential for a 15-minute access_token limited to requested scopes.Reads text others write2 findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "Exchange an authenticated bootstrap credential for a 15-minute access_token limited to..."

      What this check looks for
    • Info

      The input schema at properties.scopes.items.description mentions "credentials".

      Matched text

      inputSchema.properties.scopes.items.description: "One existing account, credentials, rights, billing, webhooks or task-acti..."

      What this check looks for
    Description, as the server wrote it
    Exchange an authenticated bootstrap credential for a 15-minute access_token limited to requested scopes. Requested scopes must be a nonempty subset of the parent's permissions. Returns the token once; child tokens stop working when the parent expires or is revoked.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • find_copiesSearch supported sources for suspected copies using a creator handle and existing text/identity matching.Reads text others write1 finding

    Why it is in this group

    • The name "find_copies" says "find": a word for reading.
    • The server marks it as not read-only (readOnlyHint: false).

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "find_copies" as not read-only (readOnlyHint: false), and its words say it reads.

      Matched text

      annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "find_copies" says "find": a word for reading.

      What this check looks for
    Description, as the server wrote it
    Search supported sources for suspected copies using a creator handle and existing text/identity matching. First create_quote(action='search', input={rights_holder_id, work_id, delegation_id, main_handle}). Results include suspected matches, searched sources, timestamps and coverage limitations; they do not determine infringement. No visual similarity search. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • investigate_urlInvestigate a supplied public URL and create a tenant-owned case.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Investigate a supplied public URL and create a tenant-owned case. First create_quote(action='investigate', input={rights_holder_id, work_id, delegation_id, url}). Results include case_id, evidence_id, page metadata, registrar/network attribution, supported contacts and retrieval limitations. CDN visibility does not prove the origin host. Investigation does not send a notice. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • monitor_caseEnroll recurring search or case-status observations within an explicit budget and expiry.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Enroll recurring search or case-status observations within an explicit budget and expiry. First create_quote(action='monitor', input={rights_holder_id, work_id, delegation_id, target_action, expires_at, per_run_limit_cents, total_limit_cents, main_handle or case_id}). target_action='search' needs main_handle; 'status' needs case_id. cadence_hours defaults to 24. Returns monitor_id; each run reserves credits and pauses when funding, authority, expiry or limits disallow work. Never schedules takedown submissions. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • pause_monitorPause future scheduled runs for an owned monitor.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Pause future scheduled runs for an owned monitor. Requires monitor scope. Preserves its original expiry and budget; an already queued operation is not cancelled. Use resume_monitor to restart within the same bounds.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • prepare_takedownPrepare a copyright notice packet for an existing case and evidence without sending it.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Prepare a copyright notice packet for an existing case and evidence without sending it. First create_quote(action='prepare', input={rights_holder_id, work_id, delegation_id, case_id, evidence_id, good_faith:true, accurate_and_authorized:true, infringement_explanation}). Optional destination must match a supported route. Returns readiness, missing fields or blockers; ready packets include packet_id and destination. Attest only claims supported by the rights holder's authority. Authenticate with the matching action scope. Reserves the quoted maximum from prepaid credits and returns operation_id; poll get_operation for results. Sandbox execution is simulated.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • replay_eventRequeue a known event to an owned active webhook while preserving its original event ID.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Requeue a known event to an owned active webhook while preserving its original event ID. Requires webhooks scope; at most ten explicit replays per delivery. The receiver must deduplicate repeated IDs. Does not rerun the underlying operation or charge it again.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • resume_monitorResume an owned monitor within its existing expiry and total budget, optionally replacing its delegation.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Resume an owned monitor within its existing expiry and total budget, optionally replacing its delegation. Requires monitor scope and current authority; expired monitors cannot resume. Does not expand spending ceilings or grant rights.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • rotate_credentialReplace the current bootstrap credential and invalidate it and its child tokens.Reads text others write1 finding

    Why it is in this group

    • The description's first verb is "replace": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "Replace the current bootstrap credential and invalidate it and its child tokens...."

      What this check looks for
    Description, as the server wrote it
    Replace the current bootstrap credential and invalidate it and its child tokens. Requires credentials scope and a bootstrap credential, not a short-lived token. Returns the new credential once; securely store it before making subsequent calls.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • set_spending_policySet organization and current-credential daily spending ceilings in USD cents.Reads text others write2 findings

    Why it is in this group

    • The name "set_spending_policy" says "set": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "Set organization and current-credential daily spending ceilings in USD cents. R..."

      What this check looks for
    • Info

      The input schema at properties.body.description mentions "credential".

      Matched text

      inputSchema.properties.body.description: "Daily organization and credential ceilings plus funding_policy (null by d..."

      What this check looks for
    Description, as the server wrote it
    Set organization and current-credential daily spending ceilings in USD cents. Requires billing scope. Set funding_policy=null to keep automatic replenishment off; enabling it requires an explicit amount, threshold, daily maximum, future expiry and mandate reference. Returns applied limits and automatic_funding state; does not itself purchase credits.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
Only reads 11Show 11 read-only tools
  • get_accountRead this credential's organization_id, principal_id, sandbox/live environment and account status.Reads private dataReads text others write1 finding

    Why it is in this group

    • The name "get_account" says "get": a word for reading.

    What it can reach

    • private-dataThe description's first sentence says "read" and "credential": it reads secrets.
    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "Read this credential's organization_id, principal_id, sandbo..."

      What this check looks for
    Description, as the server wrote it
    Read this credential's organization_id, principal_id, sandbox/live environment and account status. Use to confirm which account later work affects. Requires account scope; free, without network investigation.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_balanceRead organization-bound prepaid balance_cents, reserved_cents and available_cents in USD.Reads text others writeNo findings

    Why it is in this group

    • The name "get_balance" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read organization-bound prepaid balance_cents, reserved_cents and available_cents in USD. Requires billing scope; free. Reserved funds cannot be spent by another operation; credits cannot transfer between organizations.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_caseRead stored observations and permitted details for a case owned by this organization.Reads text others writeNo findings

    Why it is in this group

    • The name "get_case" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read stored observations and permitted details for a case owned by this organization. Free; does not fetch the target again. Use check_status with a fresh status quote for a new content observation. Returns tenant-scoped case state without internal reasoning or other customer data.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_eventsRead this organization's last 100 immutable lifecycle events with stable id, type, data and createdAt.Reads text others writeNo findings

    Why it is in this group

    • The name "get_events" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read this organization's last 100 immutable lifecycle events with stable id, type, data and createdAt. Requires webhooks scope; free. Deduplicate by event ID and fetch current state when delivery is out of order. Use replay_event for a known event and active receiver.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_evidenceExport an owned immutable metadata observation with its SHA-256 hash, capturedAt and retention expiry.Reads text others writeNo findings

    Why it is in this group

    • The name "get_evidence" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Export an owned immutable metadata observation with its SHA-256 hash, capturedAt and retention expiry. Requires evidence scope; free. Returns no screenshots, media or storage paths. Use capture_evidence with a quote to create a new observation.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_monitorRead an owned monitor's action, cadence, expiry, per-run/total budget, nextRunAt, status and pauseReason.Reads text others writeNo findings

    Why it is in this group

    • The name "get_monitor" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read an owned monitor's action, cadence, expiry, per-run/total budget, nextRunAt, status and pauseReason. Requires monitor scope; free. Does not run monitoring or increase its budget.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_operationPoll an asynchronous operation by operation_id without spending credits.Reads text others writeNo findings

    Why it is in this group

    • The name "get_operation" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Poll an asynchronous operation by operation_id without spending credits. Requires the operation's action scope. Returns queued, running, succeeded, blocked or failed with result, errorCode and reserved/settled cents. Use after any task tool; blocked submission_reconciliation_required must be reconciled before retry. A succeeded submit operation does not prove content removal.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_packetRead an owned prepared notice packet, fingerprint and any stored submission receipt.Reads text others writeNo findings

    Why it is in this group

    • The name "get_packet" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read an owned prepared notice packet, fingerprint and any stored submission receipt. Requires prepare scope; free. Does not prepare or send a notice. Use prepare_takedown to create a packet and submit_takedown for authorized delivery; a stored receipt is separate from removal evidence.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_pricesRead the authenticated environment's versioned rate card and action prices in USD cents.Reads text others writeNo findings

    Why it is in this group

    • The name "get_prices" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read the authenticated environment's versioned rate card and action prices in USD cents. Requires billing scope; free. Use create_quote to bind the current maximum price to a particular action and input.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_usageRead the last 100 ledger entries for this organization, including purchases, reservations, charges and releases with operation/purchase IDs and timestamps.Reads text others writeNo findings

    Why it is in this group

    • The name "get_usage" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Read the last 100 ledger entries for this organization, including purchases, reservations, charges and releases with operation/purchase IDs and timestamps. Requires billing scope; free. Use to reconcile spending with operation results.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_webhook_deliveriesInspect the last 100 owned webhook delivery attempts, status, attempts, last HTTP status/error, deliveredAt and replayCount.Reads text others writeNo findings

    Why it is in this group

    • The name "get_webhook_deliveries" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "inspect" and "http": it reads web pages, which anyone can write.
    Description, as the server wrote it
    Inspect the last 100 owned webhook delivery attempts, status, attempts, last HTTP status/error, deliveredAt and replayCount. Requires webhooks scope; free. Receiver failure does not lose operation or case state.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

70 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 private-data tool breaks every one.

Reads private data1
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write35
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out2
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
deny to break every path
{
  "permissions": {
    "deny": [
      "mcp__agent-access__get_account"
    ]
  }
}

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:21 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
24
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
not observed
Authorization servers
not observed
Legacy HTTP+SSE transport
no
Dynamic client registration
not observed
Client ID metadata documents
not observed
  • HighThe whole server

    The endpoint served tools/list without credentials, and the list has 24 tools that change things (4 destructive, 2 irreversible, 18 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://contentsweep.ai/api/mcp: tools/list returned 35 tools to an anonymous client, including submit_takedown (destructive), revoke_authority (destructive), revoke_credential (destructive), revoke_webhook (destructive), capture_evidence (irreversible) and 19 more

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

2 warn8 info
  • The server marks "find_copies" as not read-only (readOnlyHint: false), and its words say it reads.

    Matched text

    annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "find_copies" says "find": a word for reading.

    What this check looks for
  • The server marks "check_status" as not read-only (readOnlyHint: false), and its words say it reads.

    Matched text

    annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "check_status" says "status": a word for reading.

    What this check looks for
  • The description mentions "credentials".

    Matched text

    description: "...ctivation gates apply. Returns new live credentials once. Does not purchase credits or esta..."

    What this check looks for
  • The description mentions "credential".

    Matched text

    description: "Read this credential's organization_id, principal_id, sandbo..."

    What this check looks for
  • The description mentions "credential".

    Matched text

    description: "Set organization and current-credential daily spending ceilings in USD cents. R..."

    What this check looks for
  • The input schema at properties.body.description mentions "credential".

    Matched text

    inputSchema.properties.body.description: "Daily organization and credential ceilings plus funding_policy (null by d..."

    What this check looks for
Show 4 more findings
  • The description mentions "credential".

    Matched text

    description: "Exchange an authenticated bootstrap credential for a 15-minute access_token limited to..."

    What this check looks for
  • The input schema at properties.scopes.items.description mentions "credentials".

    Matched text

    inputSchema.properties.scopes.items.description: "One existing account, credentials, rights, billing, webhooks or task-acti..."

    What this check looks for
  • The description mentions "credential".

    Matched text

    description: "Replace the current bootstrap credential and invalidate it and its child tokens...."

    What this check looks for
  • The description mentions "credential".

    Matched text

    description: "...ately revoke an owned service-principal credential by ID. Requires credentials scope. Its..."

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.

Install

Pinned to version 1.0.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http agent-access https://contentsweep.ai/api/mcp --header 'Authorization: <your-authorization>'
claude mcp add --transport http agent-access https://contentsweep.ai/api/mcp --header 'Authorization: <your-authorization>'

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__agent-access__get_account",
      "mcp__agent-access__get_prices",
      "mcp__agent-access__get_balance",
      "mcp__agent-access__get_usage",
      "mcp__agent-access__get_operation",
      "mcp__agent-access__get_case",
      "mcp__agent-access__get_evidence",
      "mcp__agent-access__get_packet",
      "mcp__agent-access__get_monitor",
      "mcp__agent-access__get_events",
      "mcp__agent-access__get_webhook_deliveries"
    ],
    "ask": [
      "mcp__agent-access__find_copies",
      "mcp__agent-access__investigate_url",
      "mcp__agent-access__capture_evidence",
      "mcp__agent-access__prepare_takedown",
      "mcp__agent-access__check_status",
      "mcp__agent-access__monitor_case",
      "mcp__agent-access__enroll_agent",
      "mcp__agent-access__enroll_live",
      "mcp__agent-access__create_quote",
      "mcp__agent-access__purchase_credits",
      "mcp__agent-access__set_spending_policy",
      "mcp__agent-access__create_rights_holder",
      "mcp__agent-access__create_work",
      "mcp__agent-access__delegate_authority",
      "mcp__agent-access__pause_monitor",
      "mcp__agent-access__resume_monitor",
      "mcp__agent-access__exchange_token",
      "mcp__agent-access__rotate_credential",
      "mcp__agent-access__create_webhook",
      "mcp__agent-access__replay_event"
    ],
    "deny": [
      "mcp__agent-access__submit_takedown",
      "mcp__agent-access__revoke_authority",
      "mcp__agent-access__revoke_credential",
      "mcp__agent-access__revoke_webhook"
    ]
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as agent-access. Review it before you use it.

Capability tags behind 35 rules
  • mcp__agent-access__find_copies untrusted-input
  • mcp__agent-access__investigate_url untrusted-input
  • mcp__agent-access__capture_evidence untrusted-input
  • mcp__agent-access__prepare_takedown untrusted-input
  • mcp__agent-access__submit_takedown untrusted-input
  • mcp__agent-access__check_status untrusted-input
  • mcp__agent-access__monitor_case untrusted-input
  • mcp__agent-access__enroll_agent untrusted-input
  • mcp__agent-access__enroll_live untrusted-input
  • mcp__agent-access__get_account private-data, untrusted-input
  • mcp__agent-access__get_prices untrusted-input
  • mcp__agent-access__get_balance untrusted-input
  • mcp__agent-access__get_usage untrusted-input
  • mcp__agent-access__create_quote untrusted-input, outbound
  • mcp__agent-access__purchase_credits untrusted-input
  • mcp__agent-access__set_spending_policy untrusted-input
  • mcp__agent-access__create_rights_holder untrusted-input
  • mcp__agent-access__create_work untrusted-input
  • mcp__agent-access__delegate_authority untrusted-input
  • mcp__agent-access__revoke_authority untrusted-input
  • mcp__agent-access__get_operation untrusted-input
  • mcp__agent-access__get_case untrusted-input
  • mcp__agent-access__get_evidence untrusted-input
  • mcp__agent-access__get_packet untrusted-input
  • mcp__agent-access__get_monitor untrusted-input
  • mcp__agent-access__pause_monitor untrusted-input
  • mcp__agent-access__resume_monitor untrusted-input
  • mcp__agent-access__exchange_token untrusted-input
  • mcp__agent-access__rotate_credential untrusted-input
  • mcp__agent-access__revoke_credential untrusted-input
  • mcp__agent-access__create_webhook untrusted-input, outbound
  • mcp__agent-access__revoke_webhook untrusted-input
  • mcp__agent-access__get_events untrusted-input
  • mcp__agent-access__get_webhook_deliveries untrusted-input
  • mcp__agent-access__replay_event untrusted-input

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/ai.contentsweep/agent-access/verify?hash=f2937b67896fbe297d10d89002ecc81d52fa9e8398211be01e170d55a0f4d6d9" | jq -e '.changed == false'

Pinned hash f2937b67896f, from Nov 11, 2026, 18:17 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/ai.contentsweep/agent-access/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for ai.contentsweep/agent-access
[![Anlyon Radar](https://radar.anlyon.com/badge/ai.contentsweep/agent-access.svg)](https://radar.anlyon.com/servers/ai.contentsweep/agent-access)

Source: MCP Registry. Synced Oct 21, 2026, 17:41 UTC.

Content Sweep MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the Content Sweep MCP server safe?

No scan can promise that. What its definitions show: 35 tools, 24 of which can change, send or delete something, and 10 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the Content Sweep MCP server delete data?

Yes. 4 tools can delete, cancel or revoke something: submit_takedown, revoke_authority, revoke_credential and revoke_webhook.

Can the Content Sweep MCP server do things that cannot be undone?

Yes. 2 tools can send, pay, publish or merge: capture_evidence and purchase_credits.

What tools does the Content Sweep MCP server have?

35 tools: 4 destructive, 2 irreversible, 18 write, 11 read. submit_takedown, revoke_authority, revoke_credential, revoke_webhook, capture_evidence, purchase_credits, find_copies, investigate_url and 27 more.

Does the Content Sweep MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the Content Sweep MCP server send data to outside services?

2 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): create_quote and create_webhook.

Has the Content Sweep MCP server changed its tools?

Not since Anlyon Radar started tracking it. Last scanned Nov 11, 2026.