registry / AI-agent control-surface risk / passbaton

passbaton

Session continuity and persistent memory for Claude Code, OpenAI Codex CLI & Google Gemini CLI. Auto-injects past context on start, hands over hot files + build status across compaction, shares one local db across all three. Every feature is opt-in via `p

This report shows the evidence for passbaton@2.6.0. 2 of 4 known versions are blocked or flagged. See all versions.

AI Security Review

scanned 13h ago · by lpm-firewall-ai

The postinstall trigger and user-level hook configuration writes establish unconsented lifecycle mutation of AI-agent control settings. This meets the policy for blocking, even though the installer preserves existing user hooks and conditionally skips absent Codex or Gemini installations.

Static reason

One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source

Trigger

Installing the package with npm, which runs its postinstall script.

Impact

Agent configuration is changed without an affirmative setup action, enabling passbaton code to run on configured session and tool events.

Mechanism

The postinstall launches the installer, which writes Claude hook settings and registers hooks in Codex or Gemini configuration when those tools are present.

Policy narrative

An npm installation automatically launches the hook installer. The installer writes passbaton hooks into the user's Claude configuration and also updates Codex or Gemini configuration when present. These hooks can run during later agent sessions, so the lifecycle performs an unconsented change to AI-agent control settings.

AI rationale

Evidence

package.jsonscripts/postinstall.cjsdist/hooks/install.js

Decision evidence

public snapshot

AI identified this as a dangerous AI-agent capability at 98.0% confidence with low false-positive risk.

Evidence for policy risk

The package manifest runs a postinstall script during npm installation.
The postinstall script launches the hook installer automatically.
The installer writes Claude hook settings and also registers hooks with Codex and Gemini when those tools are present.

Evidence against

The installer keeps existing passbaton hook entries unchanged during automatic setup and skips Codex or Gemini when their configuration directories are absent.

Behavioral surface

Source

ChildProcessCryptoEnvironmentVarsFilesystemShell

Supply chain

HighEntropyStrings

ManifestNo manifest risk signals triggered.

scanned 50 file(s), 642 KB of source

Source & flagged code

6 flagged

package.jsonView file

Lines 11-51json

11 "passbaton-hook-user-prompt": "dist/hooks/user-prompt-submit.js",

12 "passbaton-hook-post-tool": "dist/hooks/post-tool-use.js",

13 "passbaton-hook-pre-compact": "dist/hooks/pre-compact.js",

14 "passbaton-hook-session-end": "dist/hooks/session-end.js",

15 "claude-session-continuity-mcp": "dist/index.js",

16 "claude-session-hooks": "dist/hooks/install.js",

17 "claude-hook-session-start": "dist/hooks/session-start.js",

18 "claude-hook-user-prompt": "dist/hooks/user-prompt-submit.js",

19 "claude-hook-post-tool": "dist/hooks/post-tool-use.js",

20 "claude-hook-pre-compact": "dist/hooks/pre-compact.js",

21 "claude-hook-session-end": "dist/hooks/session-end.js"

22 },

23 "scripts": {

24 "build": "tsc",

25 "dev": "tsc --watch",

26 "start": "node dist/index.js",

27 "start:v2": "MCP_V2_ONLY=true node dist/index.js",

28 "test": "vitest run",

29 "test:watch": "vitest",

30 "test:coverage": "vitest run --coverage",

31 "postinstall": "node scripts/postinstall.cjs",

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg · L31

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg · L31

32 "prepublishOnly": "npm run build && npm test"

33 },

34 "keywords": [

35 "mcp",

36 "model-context-protocol",

37 "claude",

38 "claude-code",

39 "claude-code-memory",

40 "claude-code-hooks",

41 "codex",

42 "codex-cli",

43 "openai",

44 "gemini",

45 "gemini-cli",

46 "google",

47 "ai",

48 "context",

49 "memory",

50 "session",

51 "continuity",

dist/utils/resume-card.jsView file

Lines 58-98javascript

58 return fallback;

59 return Math.min(MAX_CARD_CHARS, Math.max(MIN_CARD_CHARS, Math.floor(n)));

60}

61export function readContinuityDeclaration(projectPath

dist/utils/logger.jsView file

•Runtime or CLI code links a write operation to an explicit external AI-agent control path: export function loghookerror(hook, err) { try { const candidates = [ process.env.hook_error_log, path.join(process.cwd(), '.claude', 'hook-errors.log'), path.join(process.env.home || '/tmp', '.claude', 'hook-errors.log'), ].filter(boolean); const line = `[${new date().toisostring()}] hook=${hook} ${err instanceof error ? (err.stack || err.message) : string(err)}\n`; for (const p of candidates) { try { fs.appendfilesync(p, line); return; } catch { } } } catch { } }

High

External Ai Agent Control Surface Mutation

Runtime or CLI source writes behavior-bearing configuration into a user or project AI-agent control surface.

dist/utils/logger.jsView on unpkg

dist/utils/resume-card.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

dist/utils/resume-card.js#virtual:normalized:round1View on unpkg

dist/hooks/session-end.jsView file

•matchType = previous_version_dangerous_delta matchedPackage = passbaton@2.1.0 matchedIdentity = npm:cGFzc2JhdG9u:2.1.0 similarity = 0.692 summary = stored previous version shares package body but lacks this dangerous source file

High

Previous Version Dangerous Delta

This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.

dist/hooks/session-end.jsView on unpkg

Source & flagged code

6 flagged · loading source

package.jsonView file

•scripts.postinstall = node scripts/postinstall.cjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node scripts/postinstall.cjs

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

dist/utils/resume-card.jsView file

78try { L79: raw = JSON.parse(text.replace(/^\uFEFF/, '')); L80: }

Low

Weak Crypto

Package source references weak cryptographic algorithms.

dist/utils/resume-card.jsView on unpkg · L78

Findings

4 High3 Medium5 Low

HighInstall Time Lifecycle Scriptspackage.json

HighExternal Ai Agent Control Surface Mutationdist/utils/logger.js

HighSemantic Analysis Limiteddist/utils/resume-card.js#virtual:normalized:round1

HighPrevious Version Dangerous Deltadist/hooks/session-end.js

MediumAmbiguous Install Lifecycle Scriptpackage.json

MediumEnvironment Vars

MediumStructural Risk Force Deep Review

LowNon Install Lifecycle Scripts

LowScripts Present

LowWeak Cryptodist/utils/resume-card.js

LowFilesystem

LowHigh Entropy Strings

Affected versions and remediation

This report applies to passbaton@2.6.0.

The versions list below shows the verdict for each recorded version.

Avoid installing passbaton@2.6.0. 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: 2026-10-22 11:56:10.870Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.

Versions

4 known versions

2.6.0latestshownMaliciousSoft block: AI-agent control surfaceBlockPublishedOct 22, 2026, 10:02 AM 2.2.0MaliciousSoft block: AI-agent control surfaceBlockPublishedSep 6, 2026, 7:38 PM 2.1.0Suspicious—WarnPublishedJul 15, 2026, 4:32 PM 2.0.0Suspicious—WarnPublishedJul 15, 2026, 1:19 PM

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Versions

Related packages

Other AI-agent control-surface risk packages

Runtime surface

Behavioral surface

Related packages

Other AI-agent control-surface risk packages

Source & flagged code