@theronap/agnoclast-mcp@0.9.282
Connect your AI assistant to Cortex — your org's projects, activity, gaps, and directives, scoped to you.
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe package automatically reads Codex’s ID token and sends it to Agnoclast when a confidential-brain seat is found. Although this supports account verification, it transfers a credential to a service other than the one that issued it without an explicit user request.
Static reason
No blocking static signals were detected.; previous stored version diff introduced dangerous source
Trigger
MCP server startup when the connected account has a confidential-brain seat.
Impact
Exposes an OpenAI-issued authentication token to a different service without an explicit user request.
Mechanism
Reads the ID token from ~/.codex/auth.json and posts it in the proof field to the Agnoclast API.
Attack narrative
On server startup, the package first checks whether the Agnoclast account has a confidential-brain seat. If it does, the package reads Codex’s ID token from ~/.codex/auth.json and submits it as account proof to Agnoclast. This automatically transfers an OpenAI-issued credential to another service without an explicit user request.
AI rationale
The package automatically reads Codex’s ID token and sends it to Agnoclast when a confidential-brain seat is found. Although this supports account verification, it transfers a credential to a service other than the one that issued it without an explicit user request.
Evidence
lib/confidential_session.mjslib/server.mjslib/diagnose.mjs~/.codex/auth.json
Network endpoints1
agnoclast.com/api/brains/confidential/mine
Decision evidence
public snapshotAI called this Malicious at 97.0% confidence as Malware with low false-positive risk.
Evidence for block
- The package reads Codex’s ~/.codex/auth.json and extracts its tokens.id_token.
- When the MCP server starts, if the account has a confidential-brain seat, it gathers the agent proof and sends it in a follow-up account lookup without waiting for a user request.
- The account lookup POST serializes proof to the configured Agnoclast API base.
- The default API base is https://agnoclast.com.
Evidence against
- The ID token is used as account proof for the package’s confidential-brain feature.
- The proof lookup only runs when the initial lookup finds at least one confidential-brain seat.
- The package manifest has no install lifecycle scripts.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetwork
HighEntropyStringsUrlStrings
Source & flagged code
6 flaggedlib/capture.mjsView file
1import { readFileSync, accessSync, constants, openSync, mkdirSync } from 'fs'
L2: import { spawn, execFileSync } from 'child_process'
L3: import { homedir } from 'os'
...
L21: if (!res.ok) return null
L22: const j = await res.json().catch(() => ({}))
L23: return Array.isArray(j.types) ? j.types : null
...
L35: try {
L36: if (resolve(String(cwd)) === homedir()) return 'general'
L37: } catch { /* unresolvable path — fall through to basename */ }
...
L45: //
L46: // Handles the four remote forms git emits: scp-like ssh (git@github.com:o/n.git), ssh://, https://,
L47: // git://. The HOST CHECK IS LOAD-BEARING, not cosmetic: this string becomes a clearance key, and
Low
Weak Crypto
Package source references weak cryptographic algorithms.
lib/capture.mjsView on unpkg · L1lib/editors/antigravity.mjsView file
4Trigger-reachable persistence chain: manifest.bin -> bin/cortex-mcp.mjs -> lib/editors/claude.mjs -> lib/editors/index.mjs -> lib/editors/antigravity.mjs
L4: // WatchPaths the trajectory store and runs a one-shot sync. wire() generates the two artifacts
L5: // (the sync.sh + the launchd plist) deterministically. It does NOT run `launchctl load` here
L6: // (an unverifiable system mutation) and it flags the PORTABILITY BLOCKER below.
...
L21: /** The launchd plist, home-substituted. Deterministic — golden-testable byte-for-byte. */
L22: export function renderAntigravityPlist({ home = homedir() } = {}) {
L23: const sh = join(home, '.cortex', 'antigravity-sync.sh')
...
L33: <array>
L34: <string>/bin/bash</string>
L35: <string>${sh}</string>
Findings
4 High3 Medium5 Low
HighTrigger Reachable External Ai Agent Control Surface Mutationlib/channel.mjs
HighTrigger Reachable Persistencelib/editors/antigravity.mjs
HighSemantic Analysis Limitedbin/cortex-mcp.mjs#virtual:normalized:round1
HighPrevious Version Dangerous Deltalib/confidential_session.mjs
MediumNetwork
MediumEnvironment Vars
MediumInstall Persistencelib/editors/antigravity.mjs
LowScripts Present
LowWeak Cryptolib/capture.mjs
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to @theronap/agnoclast-mcp@0.9.282.
See all versions of @theronap/agnoclast-mcp for other recorded verdicts.
- Avoid installing @theronap/agnoclast-mcp@0.9.282. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.