Strix — open-source packagesBLOCK
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
This repository is the public release surface for the open-source pieces of Strix, an execution-control system for AI agents. Everything here runs locally and needs no Strix account. The trust primitives — the verifier, the tool-gateway, and the capability packs — are MIT-licensed; the one MCP runtime adapter (@strixgov/mcp-adapter) is source-available under Elastic License 2.0 (free to use, not to resell as a competing managed service). See LICENSING_BOUNDARY.md.
It is a mirror, not the source of truth — the canonical code lives upstream in the Strix monorepo and synchronizes here at release time. See MIRROR.md for the model and CONTRIBUTING.md for how changes flow.
Packages
d2cffa894859OBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add verifier --env ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} --env CLAUDE_API_KEY=${CLAUDE_API_KEY} --env GITHUB_PERSONAL_ACCESS_TOKEN=${GITHUB_PERSONAL_ACCESS_TOKEN} --env LINEAR_API_KEY=${LINEAR_API_KEY} -- npx -y @strixgov/[email protected]{
"mcpServers": {
"verifier": {
"command": "npx",
"args": [
"-y",
"@strixgov/[email protected]"
],
"env": {
"ANTHROPIC_API_KEY": "${ANTHROPIC_API_KEY}",
"CLAUDE_API_KEY": "${CLAUDE_API_KEY}",
"GITHUB_PERSONAL_ACCESS_TOKEN": "${GITHUB_PERSONAL_ACCESS_TOKEN}",
"LINEAR_API_KEY": "${LINEAR_API_KEY}"
}
}
}
}Exposed tools (16)
7 read · 7 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
delete_file | destructive | delete a file |
delete_thing | destructive | Delete a thing |
exec_shell | write | run a shell command |
inspect_repository | read | Inspect repository metadata without modifying the repository. |
merge_production | write | Merge an approved commit to the production branch. |
mystery_tool | read | Unclassified surface |
propose_patch | write | Create a patch proposal on a remediation branch without merging it. |
read_file | read | read a file |
read_thing | read | Read a thing |
rotate_credentials | read | Rotate a production credential. |
strix_verify | read | Independently verify any Strix governance artifact by passing raw @strixgov/verifier CLI arguments. |
strix_verify_record | read | Convenience: verify a single Strix evidence record by its evidenceId. Optionally override the proof/JWKS |
strix_verify_swarm | write | Convenience: verify an agent-swarm run by its swarmRunId (re-derives the delegation-chain verdict). |
write_deal_summary | write | ... |
write_file | write | write a file |
write_thing | write | Write a thing |
Trust audit
BLOCKgrade F · trust 44/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | FAIL |
| L1 | Static analysis of the code | WARN |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (8 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (25)
xag-neg-01-payload.json
xag-neg-07-presented-payload.json
rcpt-neg-01-tampered-payload.json
actor-attestation.ts
network-signatures.ts
sign.ts
"Fetch an arbitrary URL. A read that can exfiltrate: outbound request bytes are attacker-shapable.",
out = { status: 200, body: { evidenceId: 'ev_self_check', proofUrl: 'http://127.0.0.1/proof/ev_self_check' } };return { base: `http://127.0.0.1:${server.address().port}`, close: () => server.close() };const secret = "test-secret-at-least-16-chars-long";
secret: "wrong-secret-at-least-16-chars",
secret: "test-secret-at-least-16-chars-long",
apiKey: "0123456789abcdef0123456789abcdef",
const apiKey = "0123456789abcdef0123456789abcdef";
assert.match(pem, /^-----BEGIN PRIVATE KEY-----/);
delete_file, delete_thing
mcp1-scitt-neg-01-tampered-payload.json
const out = exec ? await exec(tu.input ?? {}) : { ok: false, reason: "unknown capability" };const ROOT = path.resolve(HERE, '../../../conformance/corpus');
() => respond("../../etc/passwd", { approved: true, env: f.env }),import { canonicalizeJSON } from "../../src/canonical-json.mjs";import { buildCanonicalPayload } from "../../src/index.mjs";} from "../../src/index.mjs";
test("query / read_query / explain_query are MEDIUM READ (exfiltration risk)", () => {test("Postgres reads are intentionally MEDIUM (exfiltration discipline)", () => {Gates applied: no_behavioural_pass.
d2cffa894859full audit observations/trust-audit/mcp-server/strixgov__strix.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | d2cffa894859 | BLOCK | F | 44 | first audit |
Questions
What tools does @strixgov/mcp-proxy expose?
16 in total: 7 read-only, 7 that write, and 2 that can delete or overwrite (delete_file, delete_thing). Every one is listed on this page with its risk.
Is @strixgov/mcp-proxy safe to connect to an agent?
No — not without reading the findings first. The audit graded it F (44/100) and found 3 critical or high issues in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does @strixgov/mcp-proxy need?
It reads ANTHROPIC_API_KEY, CLAUDE_API_KEY, GITHUB_PERSONAL_ACCESS_TOKEN, LINEAR_API_KEY, NOTION_TOKEN, SLACK_MCP_XOXP_TOKEN, STRIX_AGENT_KEY_REGISTRY, STRIX_API_KEY, STRIX_CLAUDE_KEY_DIR, STRIX_DECISION_TOKEN_SECRET, STRIX_TEST_TOKEN and STRIX_TEST_TOKEN_2 from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does @strixgov/mcp-proxy run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as @strixgov/verifier at 1.24.0.
How current is this page?
The grade is for one exact copy of the source (d2cffa894859), read on 2026-10-12. The repository is watched and re-audited when it changes.