ExitProof MCP server
com.davisvillelabs/exitproofdavisvillelabs.comstreamable-httpv0.1.0Scanned
Reversibility intelligence and durable exit evidence for AI agents before real-world commitments.
6 tools: 3 do something that can't be undone, 1 change things and 2 only read.
- Reads private data · 0
- Reads text others write · 3
- Can send data out · 1
- 17 findings
The remote endpoint did not answer the last look. Remote posture
Anlyon alternative
4 of these tools change something. Run those as governed actions.
Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.
Anlyon Radar does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.
Import an OpenAPI spec →What the ExitProof MCP server can do
The ExitProof MCP server is published by davisvillelabs.com, a domain the MCP Registry verified. It has 6 tools. It can do things that cannot be undone, such as sending, paying or publishing (check_reversibility, check_agent_purchase and create_exit_manifest) and create or change data (1 tool: build_exit_pack), but cannot delete data. 2 tools only read.
3 tools bring in text other people can write (check_agent_purchase, create_exit_manifest and build_exit_pack); 1 tool can send data out of your system (check_agent_purchase).
The fixed checks matched 17 findings (credentials-mention and annotation-conflict). Each one comes with the text it matched.
Last scanned Oct 21, 2026, version 0.1.0.
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.
6 tools
Can't be undone 3Claude Code: ask
check_agent_purchaseInspect one autonomous HTTP purchase before the calling agent pays it.Reads text others writeCan send data out5 findings
Why it is in this group
- The name "check_agent_purchase" 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.
- outboundThe input takes a URL (input.target_url): a request to a URL can carry data to whoever runs it.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...gnals. ExitProof never forwards payment credentials and never sends a live POST, PUT, PATCH..."
- Info
The input schema at description mentions "credential".
What this check looks forMatched text
inputSchema.description: "...P purchase. ExitProof either performs a credential-free GET/HEAD probe or analyzes a 402 c..."
- Info
The input schema at properties.target_url.description mentions "Credentials".
What this check looks forMatched text
inputSchema.properties.target_url.description: "...dpoint the agent is considering paying. Credentials, localhost, private-address literals, a..."
- Info
The input schema at properties.live_probe.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.live_probe.description: "...ether ExitProof may make one anonymous, credential-free request to target_url. Defaults tr..."
- Info
The input schema at properties.observed_response.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.observed_response.description: "...challenge headers only, never a payment credential."
Description, as the server wrote it Inspect one autonomous HTTP purchase before the calling agent pays it. This $0.01 machine-payment operation validates a real 402 challenge, compares advertised price/protocol/network/asset/payment destination when supplied, checks x402 v2 exact resource binding when available, and surfaces retry or settled-delivery recovery signals. ExitProof never forwards payment credentials and never sends a live POST, PUT, PATCH, or DELETE probe; for those methods the caller supplies an already-observed 402 challenge. Conflicting, malformed, inaccessible, non-402, or unsafe targets are rejected before ExitProof issues its own payment challenge. Use check_reversibility instead for cancellation/refund terms on a real-world commitment, create_exit_manifest to preserve an eligible exit path, and build_exit_pack when an exit is actually needed. A successful check is not a recommendation, merchant-identity guarantee, delivery guarantee, legal opinion, or authorization to spend.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
check_reversibilityEvaluate one commitment’s cancellation/refund reversibility from transaction-specific checkout terms plus maintained policy evidence, without charging or creating durable state.3 findings
Why it is in this group
- The description's first verb is "evaluate": a word for a change.
- The input carries an amount (input.amount) and a currency, so it moves money.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Info
The input schema at properties.merchant.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.merchant.description: "...erlays; do not include account numbers, credentials, or secrets."
- Info
The input schema at properties.checkout_terms.properties.source_url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.checkout_terms.properties.source_url.description: "...ed transaction terms. ExitProof removes credentials, query parameters, and fragments before..."
- Warn
The server marks "check_reversibility" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The description's first verb is "evaluate": a word for a change.
Description, as the server wrote it Evaluate one commitment’s cancellation/refund reversibility from transaction-specific checkout terms plus maintained policy evidence, without charging or creating durable state. Use before a purchase, booking, subscription, rental, service, ticket, digital purchase, or deposit when exit deadlines or loss matter; use list_supported_policies only for maintained-overlay coverage, and create_exit_manifest only after paidExitManifestAvailable=true. Do not use for purchase desirability, legal advice, chargeback decisions, or executing a cancellation. Supply checkout_terms whenever available because they control the transaction-specific assessment; expected_commitment_at moves deadline evaluation forward when the commitment is future-dated, and missing or conflicting evidence can return assessment=unknown rather than inventing an exit right.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
create_exit_manifestPersist a privacy-minimized Exit Manifest for an eligible commitment, then return its manifest_id and secret access_token.Reads text others write3 findings
Why it is in this group
- The name "create_exit_manifest" says "create": a word for a change.
- The input carries an amount (input.amount) and a currency, so it moves money.
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".
What this check looks forMatched text
description: "...cal settled retry with the same payment credential recovers the same manifest without a se..."
- Info
The input schema at properties.merchant.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.merchant.description: "...erlays; do not include account numbers, credentials, or secrets."
- Info
The input schema at properties.checkout_terms.properties.source_url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.checkout_terms.properties.source_url.description: "...ed transaction terms. ExitProof removes credentials, query parameters, and fragments before..."
Description, as the server wrote it Persist a privacy-minimized Exit Manifest for an eligible commitment, then return its manifest_id and secret access_token. Use only after check_reversibility says paidExitManifestAvailable=true; pass the exact commitment facts to preserve, with transaction-specific checkout_terms controlling the assessment when supplied. Use check_reversibility for evaluation only, or get_exit_manifest when a manifest already exists. This state-changing $0.25 stablecoin MPP operation checks eligibility before any payment challenge; a fresh authorization creates durable state, while an identical settled retry with the same payment credential recovers the same manifest without a second settlement. The plaintext access token is returned to the caller but stored only as a SHA-256 hash.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
Changes things 1Claude Code: ask
build_exit_packBuild a bounded cancellation/refund action pack from an existing actionable Exit Manifest.Reads text others write3 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".
What this check looks forMatched text
description: "...est_id and access_token together as the credential pair returned by create_exit_manifest o..."
- Info
The input schema at description mentions "Credentials".
What this check looks forMatched text
inputSchema.description: "Credentials for one previously created Exit Manifes..."
- Info
The input schema at properties.access_token.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.access_token.description: "...uired with manifest_id; treat it like a credential and do not log or expose it."
Description, as the server wrote it Build a bounded cancellation/refund action pack from an existing actionable Exit Manifest. Pass manifest_id and access_token together as the credential pair returned by create_exit_manifest or settlement recovery; use get_exit_manifest for retrieval only, or check_reversibility then create_exit_manifest when no manifest exists. This $1.00 MPP operation prefers Tempo stablecoin when available, retains Stripe card/link as a compatible fallback, and can charge but never contacts the merchant, executes a cancellation, files a chargeback, or guarantees success. Invalid or mismatched credentials and non-actionable manifests are rejected before payment; an identical settled retry with the same payment credential avoids a second settlement, while generatedAt and deadlineStatus can reflect the current time.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
Only reads 2Show 2 read-only tools
get_exit_manifestRetrieve one previously created Exit Manifest without charging or changing stored state.3 findings
Why it is in this group
- The name "get_exit_manifest" says "get": a word for reading.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...idence; missing, expired, or mismatched credentials fail without modifying the manifest."
- Info
The input schema at description mentions "Credentials".
What this check looks forMatched text
inputSchema.description: "Credentials for one previously created Exit Manifes..."
- Info
The input schema at properties.access_token.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.access_token.description: "...uired with manifest_id; treat it like a credential and do not log or expose it."
Description, as the server wrote it Retrieve one previously created Exit Manifest without charging or changing stored state. Pass manifest_id and access_token together: the ID selects the record and the secret token authorizes access to that same record, so neither value is sufficient alone. Use this for preserved evidence and terms; use build_exit_pack for an actionable cancellation/refund plan, or create_exit_manifest when no manifest exists. This free, repeatable read does not refresh, re-evaluate, or extend evidence; missing, expired, or mismatched credentials fail without modifying the manifest.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
list_supported_policiesList the maintained merchant policy overlays and snapshot version/date ExitProof can add to transaction-specific evidence.No findings
Why it is in this group
- The name "list_supported_policies" says "list": a word for reading.
Description, as the server wrote it List the maintained merchant policy overlays and snapshot version/date ExitProof can add to transaction-specific evidence. Use this only to inspect maintained coverage; use check_reversibility to evaluate an actual commitment, including an unlisted merchant when transaction-specific checkout_terms are available. An absent merchant means no maintained overlay, not that no cancellation, refund, contractual, or legal right exists. This is a free read-only snapshot: it creates no durable state, issues no payment challenge, and does not fetch or verify a merchant’s live policy at call time.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
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.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write3
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out1
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
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.
Looked at Nov 20, 2026, 19:22 UTC, unchanged since Oct 29, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- no
- HTTP status of an anonymous initialize
- not observed
- tools/list served without credentials
- no
- Tools that change things, listed without credentials
- 0
- Protocol version
- not observed
- 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
The probe stopped: The server redirected a POST to a different request.
Posture changes
- Oct 29, 2026
- Reachable: yes to no
- HTTP status of an anonymous initialize: 200 to none
- tools/list served without credentials: yes to no
- Tools that change things, listed without credentials: 4 to 0
- Protocol version: 2025-06-18 to none
Findings
Matches of fixed checks, with the text they matched.
The server marks "check_reversibility" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The description's first verb is "evaluate": a word for a change.
The input schema at properties.merchant.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.merchant.description: "...erlays; do not include account numbers, credentials, or secrets."
The input schema at properties.checkout_terms.properties.source_url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.checkout_terms.properties.source_url.description: "...ed transaction terms. ExitProof removes credentials, query parameters, and fragments before..."
The description mentions "credentials".
What this check looks forMatched text
description: "...gnals. ExitProof never forwards payment credentials and never sends a live POST, PUT, PATCH..."
The input schema at description mentions "credential".
What this check looks forMatched text
inputSchema.description: "...P purchase. ExitProof either performs a credential-free GET/HEAD probe or analyzes a 402 c..."
The input schema at properties.target_url.description mentions "Credentials".
What this check looks forMatched text
inputSchema.properties.target_url.description: "...dpoint the agent is considering paying. Credentials, localhost, private-address literals, a..."
Show 11 more findingsShow fewer
The input schema at properties.live_probe.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.live_probe.description: "...ether ExitProof may make one anonymous, credential-free request to target_url. Defaults tr..."
The input schema at properties.observed_response.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.observed_response.description: "...challenge headers only, never a payment credential."
The description mentions "credential".
What this check looks forMatched text
description: "...cal settled retry with the same payment credential recovers the same manifest without a se..."
The input schema at properties.merchant.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.merchant.description: "...erlays; do not include account numbers, credentials, or secrets."
The input schema at properties.checkout_terms.properties.source_url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.checkout_terms.properties.source_url.description: "...ed transaction terms. ExitProof removes credentials, query parameters, and fragments before..."
The description mentions "credentials".
What this check looks forMatched text
description: "...idence; missing, expired, or mismatched credentials fail without modifying the manifest."
The input schema at description mentions "Credentials".
What this check looks forMatched text
inputSchema.description: "Credentials for one previously created Exit Manifes..."
The input schema at properties.access_token.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.access_token.description: "...uired with manifest_id; treat it like a credential and do not log or expose it."
- Infobuild_exit_pack
The description mentions "credential".
What this check looks forMatched text
description: "...est_id and access_token together as the credential pair returned by create_exit_manifest o..."
- Infobuild_exit_pack
The input schema at description mentions "Credentials".
What this check looks forMatched text
inputSchema.description: "Credentials for one previously created Exit Manifes..."
- Infobuild_exit_pack
The input schema at properties.access_token.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.access_token.description: "...uired with manifest_id; treat it like a credential and do not log or expose it."
History
Changes to its tool definitions and to its remote endpoint, newest first.
PostureRemote posture changed. Reachable: yes -> no; HTTP status of an anonymous initialize: 200 -> none; tools/list served without credentials: yes -> no (and 2 more).View diff
- Reachable: yes to no
- HTTP status of an anonymous initialize: 200 to none
- tools/list served without credentials: yes to no
- Tools that change things, listed without credentials: 4 to 0
- Protocol version: 2025-06-18 to none
Install
Pinned to version 0.1.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http exitproof https://exitproof.davisvillelabs.com/mcpclaude mcp add --transport http exitproof https://exitproof.davisvillelabs.com/mcpRun 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.
{
"permissions": {
"allow": [
"mcp__exitproof__get_exit_manifest",
"mcp__exitproof__list_supported_policies"
],
"ask": [
"mcp__exitproof__check_reversibility",
"mcp__exitproof__check_agent_purchase",
"mcp__exitproof__create_exit_manifest",
"mcp__exitproof__build_exit_pack"
],
"deny": []
}
}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 exitproof. Review it before you use it.
Capability tags behind 3 rules
- mcp__exitproof__check_agent_purchase untrusted-input, outbound
- mcp__exitproof__create_exit_manifest untrusted-input
- mcp__exitproof__build_exit_pack 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/com.davisvillelabs/exitproof/verify?hash=5360ac3ef639899b7a728df4b4c418ccea205dbb6515686e0018584ebdd6d206" | jq -e '.changed == false'Pinned hash 5360ac3ef639, from Oct 21, 2026, 19:10 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/com.davisvillelabs/exitproof/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/com.davisvillelabs/exitproof)Source: MCP Registry. Synced Oct 21, 2026, 17:53 UTC.
ExitProof MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the ExitProof MCP server safe?
- No scan can promise that. What its definitions show: 6 tools, 4 of which can change, send or delete something, and 17 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 ExitProof MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the ExitProof MCP server do things that cannot be undone?
- Yes. 3 tools can send, pay, publish or merge: check_reversibility, check_agent_purchase and create_exit_manifest.
What tools does the ExitProof MCP server have?
- 6 tools: 3 irreversible, 1 write, 2 read. check_reversibility, check_agent_purchase, create_exit_manifest, build_exit_pack, get_exit_manifest and list_supported_policies.
Does the ExitProof MCP server need sign-in or OAuth?
- Its remote endpoint did not answer the last anonymous look.
Does the ExitProof MCP server send data to outside services?
- 1 tool can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): check_agent_purchase.
Has the ExitProof MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Oct 21, 2026.