Borrowing it
Nothing to install: this file belongs to davidnichols-ops/trustcard. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/davidnichols-ops/trustcard/master/AGENTS.mdgit clone --depth 1 https://github.com/davidnichols-ops/trustcardWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/davidnichols-ops/trustcard/agents-md)<a href="https://agentmods.dev/instructions/davidnichols-ops/trustcard/agents-md"><img src="https://agentmods.dev/badge/instructions/davidnichols-ops/trustcard/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.02467 | $0.02467 |
| Opus 5 | $0.01234 | $0.01234 |
| Sonnet 5 | $0.00493 | $0.00493 |
| Haiku 4.5 | $0.00247 | $0.00247 |
Grade A, and why
trustcard AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
trustcard — repo knowledge for agents
What this is
trustcard is cryptographic trust infrastructure for MCP servers, and the start of a protocol-neutral trust substrate for executable capabilities. v3.0.3 builds on v1 identity/provenance/v2 policy/receipts with three new layers:
- Behavioral verification (
lib/behavior.js) — sandboxed, deterministic probing that can fail a server whose static contract is unchanged. - Evidence substrate (
lib/evidence.js,lib/evidence-store.js) — content-addressed, immutable observations. - HTTP/SSE transport parity (
lib/client-http.js) — experimental streamable-http and SSE client.
Read docs/INDEX.md first, then docs/SECURITY-MODEL.md, docs/SPEC.md, and docs/BEHAVIOR.md.
Commands
- Test:
npm test→node --test test/*.test.js(529 tests, all should pass)- IMPORTANT: the
test/*.test.jsglob is required. Barenode --testalso matchestest/helpers.jsand the fixture servers, which hang the runner (child processes keep stdio open). Don't "simplify" the glob away. - The glob must be unquoted in the
package.jsonscript so the shell expands it before Node receives the paths; a quoted glob becomes a literal path.
- IMPORTANT: the
- CLI:
node bin/mcp-trustcard.js <subcommand>(keygen/manifest/sign/verify/ diff/pin/pins/fingerprint/scan/gen-manifest/inspect/auth-issue/auth-verify/ behavior/evidence) - Also
bin/mcp-proxy.jsandbin/mcp-http-proxy.jsfor call-time enforcement. - No runtime deps. Pure Node stdlib. Don't add dependencies.
Architecture map (lib/)
canon.js— RFC 8785 (JCS) canonicalization. Every digest depends on this being byte-exact. Numbers use ECMAScript shortest round-trip withe+n/e-nexponents, no-0; keys sorted by UTF-16 code units. RFC test vectors intest/canon.test.js.hash.js—digest(value) = "sha256:"+base64url(SHA256(JCS(value))).signingPayload()excludes BOTHsignatureandmanifestDigest.identity.js— semantic projection for tool/server digests. Semantic: name, description, inputSchema, outputSchema, annotations{title, readOnlyHint, destructiveHint, idempotentHint, openWorldHint}, execution. Volatile (excluded): top-level title, icons, tags,_meta.diff.js— five-level taxonomy NONE<SYNTACTIC<NON_BREAKING<ANNOTATION_DOWNGRADE<PERMISSION_CHANGE<BREAKING. Compatible = ≤NON_BREAKING.change.js— multi-axis change vector (interface/permission/implementation/ provenance) for descriptors.trust.js— state machine UNKNOWN→OBSERVED→PINNED→MISMATCH/SUSPECT→REVOKED. REVOKED is sticky.trustLevel()projects onto TRUSTED/VERIFIED/OBSERVED/UNTRUSTED/REVOKED.provenance.js— manifest build/sign/verify. Ed25519 via node:crypto. Signed manifests carryexpiresAt;verifyManifestrejects expired ones.manifest.js— proxy-enforcement manifest.buildManifestincludesexpiresAt(default 90 days);checkCallblocks all calls when expired.--allow-toolrecordsmanualOverride: true.--require-scopesaddsrequiredScopesper tool.pin.js— TOFU pin store (servers + publisher keys), atomic writes, fail-closed on corrupt file.session.js— live connection: negotiates protocol, verifies handshake binding, subscribes tonotifications/tools/list_changed→ re-diff.guard.js— the enforcement gate.wrapClient/session.callroute everytools/callthroughguard.authorizeCall. Modes: enforce/audit/off. Emits signed, chained receipts when given areceiptKey.middleware.js—wrapClient(rawClient, {guard, session}); re-verifies toolset digest on everytools/list.policy.js— Gate 2 per-invocation authorization: composable rule predicates (allowTools,denyTools,constrainArg,forbidArg,restrictToolToEnvironments,requireApprovalForDestructive,requireScopes) +ScopedDecisions. Verdict precedence: deny > require-approval > allow. Rule predicates that throw fail closed.defaultVerdictdefaults toallow; set todenyfor default-deny.auth.js— Per-agent auth scope enforcement.DevIssuer(HMAC-SHA256 JWT-like tokens),IdpIntrospector(RFC 7662 OAuth 2.1 token introspection),TokenValidator(local first, then introspection),scopeSatisfies(wildcards*andprefix:*),extractToken/stripAuth.rotation.js— old-key-signs-new-key rotation certificates + self-signed revocation certificates.behavior.js— v3. Behavioral verification engine. ExportsSandboxRuntime,InputGenerator,OutputComparator,BehaviorEngine,BehaviorReport,BehaviorFinding,RegressionCorpus,ReferenceObservation. Key invariant: static identity can pass while behavioral verification fails.evidence.js/evidence-store.js/evidence-adapters.js— v3. Content-addressed evidence records and local store. Records are immutable; corrections usesupersedes.EVIDENCE_LAYERSis the predicate registry.evidence-predicates.js— predicate and layer vocabulary for evidence.existence.js— existence probes for npm and GitHub repos (evidence layer).observe.js,fingerprint.js,receipts.js,report.js,checks.js— scan, identity card, reproducibility, rendering.danger-detector.js— three-engine fusion: heuristic, semantic (TF-IDF), injection markers.descriptor.js— protocol-neutral capability descriptor.interfaceDigest()is byte-equal totoolDigest().implementationIdentityis typed (npm-dist/source/unresolved).buildDescriptor/signDescriptor/verifyDescriptorand manifest⇄descriptor adapters.mcp-server.js— minimal framework for well-behaved MCP servers.examples/bruce-lee/is the reference server.client-http.js— v3. MCPstreamable-httpand SSE client. Experimental;McpHttpClientmatchesMcpStdioClientinterface.redact.js— auth-token and sensitive-value redaction helpers.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 161 lines · 2,467 tokens per session scan A cad671ded43a
trustcard AGENTS.md is an instructions file published in the GitHub repository davidnichols-ops/trustcard (4 stars, last pushed 17d ago), licensed MIT. It adds 2,467 tokens to every session, about $0.0123 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.