Loading npm security reports…

mandrel@2.73.0: AI-agent control-surface risk | LPM Firewall

registry / mandrel / 2.73.0

mandrel@2.73.0

Claude Code-first opinionated workflow framework: instructions, skills, rules, and SDLC workflows that govern AI coding assistants.

AI Security Review

scanned 6d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installation automatically overwrites and prunes the consumer's shared .agents directory. The operation does not verify that affected instructions, skills, or rules belong to mandrel.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review

Trigger

npm installation with lifecycle scripts enabled in a consumer project.

Impact

Existing agent instructions and capabilities can be replaced or removed without consent, changing the consumer's agent control surface.

Mechanism

The postinstall hook selects INIT_CWD as the consumer root, copies the package payload into .agents, and deletes destination files missing from that payload.

Policy narrative

The published postinstall hook automatically synchronizes into the consumer project. Synchronization treats the shared .agents tree as managed content, overwriting matching files and deleting other files without an ownership ledger check. Limited local-file exemptions do not protect ordinary pre-existing instructions, rules, or skills. This crosses the boundary from first-party extension setup into unconsented mutation of a broad agent control surface.

AI rationale

Source establishes automatic, unconsented replacement and deletion across a shared agent configuration directory. This meets the blocking policy for broad install-time agent control-surface mutation, despite the package-aligned workflow purpose. Product guard normalized a concrete AI-agent control hijack publish_block to the blockable dangerous-capability shape.

Evidence

package.jsonbin/postinstall.jslib/cli/sync.js.agents

Decision evidence

Source & flagged code

7 flagged · loading source

package.jsonView file

•scripts.postinstall = node bin/postinstall.js mandrel sync

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node bin/postinstall.js mandrel sync

Medium

Ambiguous Install Lifecycle Script

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

package.jsonView on unpkg

.agents/scripts/lib/qa/redact-evidence.jsView file

17patternName = generic_password severity = medium line = 17

Medium

Secret Pattern

Package contains a possible secret pattern.

.agents/scripts/lib/qa/redact-evidence.jsView on unpkg · L17

bin/mandrel.jsView file

101function installedVersion() { L102: const req = createRequire(import.meta.url); L103: const manifest = req('../package.json');

Medium

Dynamic Require

Package source references dynamic require/import behavior.

bin/mandrel.jsView on unpkg · L101

bin/postinstall.jsView file

1Install-time AI-agent control hijack evidence: Install-time helper chain writes an AI-agent control surface: L1: #!/usr/bin/env node L2: // bin/postinstall.js — best-effort `.agents/` materializer on install. L3: ... L7: * doctor-detected "not materialized" state instead of failing the install. L8: * In the framework's own source checkout it no-ops, since `.agents/` there is L9: * the committed product and a sync would clobber it. ... L19: const HINT = L20: 'mandrel: could not materialize ./.agents/ during install — run `mandrel sync` to finish setup.\n'; L21: Referenced control path from bin/postinstall.js: L1: #!/usr/bin/env node L2: // bin/postinstall.js — best-effort `.agents/` materializer on install. L3: ... L7: * doctor-detected "not materialized" state instead of failing the ins

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

bin/postinstall.jsView on unpkg · L1

lib/cli/init.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 2

High

Semantic Analysis Limited

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

lib/cli/init.js#virtual:normalized:round1View on unpkg

lib/cli/update.jsView file

•matchType = normalized_sha256 matchedPackage = mandrel@2.60.0 matchedPath = lib/cli/update.js matchedIdentity = npm:bWFuZHJlbA:2.60.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

lib/cli/update.jsView on unpkg

Source & flagged code

7 flagged

package.jsonView file

Lines 51-91json

51 "cyclomatic:update": "node .agents/scripts/check-cyclomatic.js --update",

52 "dead-exports:update": "node scripts/update-dead-exports-baseline.js && node scripts/update-dead-exports-baseline.js --production",

53 "check:knip-entries": "node scripts/check-knip-entries.js",

54 "validator:gen": "node scripts/check-generated-validator.js",

55 "validator:check": "node scripts/check-generated-validator.js --check",

56 "config:ceiling": "node scripts/check-agentrc-key-ceiling.js",

57 "check:comment-policy": "node scripts/check-comment-policy.js",

58 "check:schema-refs": "node scripts/check-schema-references.js",

59 "check:workflow-citations": "node scripts/check-workflow-citations.js",

60 "crap:check": "node .agents/scripts/check-baselines.js --gate crap",

61 "crap:update": "node .agents/scripts/update-crap-baseline.js",

62 "crap:reanchor": "node .agents/scripts/update-crap-baseline.js --full-scope",

63 "duplication:check": "node .agents/scripts/check-baselines.js --gate duplication",

64 "duplication:update": "node .agents/scripts/update-duplication-baseline.js",

65 "duplication:reanchor": "node .agents/scripts/update-duplication-baseline.js --full-scope",

66 "quality:preview": "node .agents/scripts/check-dead-exports.js && node .agents/scripts/quality-preview.js",

67 "quality:watch": "node .agents/scripts/quality-watch.js",

68 "sync:commands": "node bin/mandrel.js sync-commands",

69 "sync:agents": "node .agents/scripts/sync-claude-agents.js",

70 "prepare": "husky && npm run sync:commands && npm run sync:agents && npm run baselines:merge-driver",

71 "postinstall": "node bin/postinstall.js mandrel sync"

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg · L71

Medium

Ambiguous Install Lifecycle Script

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

package.jsonView on unpkg · L71

72 },

73 "repository": {

74 "type": "git",

75 "url": "git+https://github.com/dsj1984/mandrel.git"

76 },

77 "keywords": [

78 "claude-code",

79 "ai-agents",

80 "sdlc",

81 "workflow",

82 "orchestration",

83 "agentic"

84 ],

85 "author": "",

86 "license": "MIT",

87 "bin": {

88 "mandrel": "bin/mandrel.js"

89 },

90 "type": "module",

91 "bugs": {

.agents/scripts/lib/qa/redact-evidence.jsView file

Lines 1-37javascript

1/**

2 * redact-evidence.js — deterministic secrets/PII scrubber run before QA

3 * evidence is persisted or posted (security baseline § Data Leakage).

4 * Idempotent (placeholders match no rule) and pass-through for benign text.

5 */

6

7/**

8 * Per-rule placeholders, free of anything a pattern matches (`@`, `=`, long

9 * token or digit runs) so the pass is a fixed point.

10 *

11 *

bin/mandrel.jsView file

Lines 81-121javascript

81 ],

82 [

83 'explain',

84 {

85 description: 'print resolved config values and their sources',

86 knownFlags: new Set(['--json']),

87 },

88 ],

bin/postinstall.jsView file

Lines 1-21javascript

1#!/usr/bin/env node

2// bin/postinstall.js — best-effort `.agents/` materializer on install.

3

4/**

5 * Runs `mandrel sync` on install. Best-effort by contract: always exits 0 so

6 * `--ignore-scripts`, sandboxed CI, or a copy error degrade to the

7 * doctor-detected "not materialized" state instead of failing the install.

8 * In the framework's own source checkout it no-ops, since `.agents/` there is

9 * the committed product and a sync would clobber it.

10 */

11

lib/cli/init.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 2

High

Semantic Analysis Limited

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

lib/cli/init.js#virtual:normalized:round1View on unpkg

lib/cli/update.jsView file

•matchType = normalized_sha256 matchedPackage = mandrel@2.60.0 matchedPath = lib/cli/update.js matchedIdentity = npm:bWFuZHJlbA:2.60.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

lib/cli/update.jsView on unpkg