Loading npm security reports…

agentic-cortex@7.2.0: AI-agent control-surface risk | LPM Firewall

registry / agentic-cortex / 7.2.0

agentic-cortex@7.2.0

Persistent self-improving agent memory with tree-of-thoughts reasoning: MCTS/beam search, PRM step verification, self-consistency voting, s1 budget forcing, REPL program-aided reasoning, reflexion self-correction, failure classification, experience replay

AI Security Review

scanned 25d ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the package automatically alters the consumer project's AI-agent configuration and instruction files. It also installs Git hooks that invoke the package after common repository events.

Static reason

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

Trigger

npm installation of the package as a dependency

Impact

The package makes its local server and instructions active in multiple agent tools without an explicit setup command, then persists execution through repository hooks.

Mechanism

Post-install modification of AI-agent controls and Git hooks

Policy narrative

On installation, the post-install script resolves the consuming project directory. It writes agent server configuration for Claude Code, Cursor, and OpenCode, arranges for an instruction file to be loaded, and creates additional agent guidance. It also creates four Git hooks that run the package's commands after commits, checkouts, merges, and rewrites. These automatic changes affect foreign project and agent control surfaces without an explicit user setup action.

AI rationale

This is concrete, automatic post-install mutation of a consumer project's broad AI-agent control surface, combined with persistent Git-hook execution. That meets the install-control-surface block policy even without confirmed network exfiltration.

Evidence

package.jsonscripts/auto-setup.jsscripts/create-discovery-files.jsknowledge.md.mcp.json.cursor/mcp.jsonopencode.jsonAGENTS.md

Source & flagged code

17 flagged · loading source

package.jsonView file

•scripts.postinstall = node scripts/auto-setup.js

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 6

High

Semantic Analysis Limited

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

package.jsonView on unpkg

src/core/hooks.jsView file

185if (process.env.AGENTIC_CORTEX_HOOKS_JS !== '1') return false; L186: try { return new Function('obs', 'ctx', value)(observation, context); } catch { return false; } L187: default:

Low

Eval

Package source references a known benign dynamic code generation pattern.

src/core/hooks.jsView on unpkg · L185

scripts/create-discovery-files.jsView file

5Install-time code directly mutates a foreign AI-agent control surface: L5: * Creates files that various AI coding agents automatically read at session start: L6: * - AGENTS.md (OpenCode, Claude Code — canonical MCP-first discovery) L7: * - .claude/CLAUDE.md (Claude Code) L8: * - .cursor/rules/agentic-cortex.mdc (Cursor, alwaysApply: true) L9: * - .opencode/agentic-cortex.md (OpenCode) L10: * L11: * Also creates MCP server configuration files: L12: * - .mcp.json (Claude Code) L13: * - .cursor/mcp.json (Cursor) L14: * - opencode.json (OpenCode) L15: * ... L95: * Idempotent — skips if agentic-cortex already configured. Write operation from scripts/auto-setup.js: L89: if (fs.existsSync(templatePath)) { L90: fs.copyFileSync(templatePath, knowledgePath); L91: c

Critical

Ai Agent Control Hijack

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

scripts/create-discovery-files.jsView on unpkg · L5

•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation. , 'agents.md'); const result = writesection(agentspath, agentsmdcontent); if (result !== 'unchanged') created.push('agents.md'); } catch (err) { console.error('[agentic-cortex] agents.md skipped: ' + err.message); } try { const claudedir = path.join(projectdir, '.claude'); if (!fs.existssync(claudedir)) fs.mkdirsync(claudedir, { recursive: true }); const claudemdcontent = [ '<agentic_cortex>', 'agentic-cortex is installed. persistent, self-improving memory across sessions.', 'mcp server is configured in .mcp.json — ' + mcp_tool_count + ' too

High

Trigger Reachable External Ai Agent Control Surface Mutation

Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.

scripts/cortex-hook.shView file

•path = scripts/cortex-hook.sh kind = build_helper sizeBytes = 7908 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

scripts/cortex-hook.shView on unpkg

scripts/import-community-knowledge.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = scripts/import-community-knowledge.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

scripts/import-community-knowledge.jsView on unpkg

scripts/seed-memory-repo.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = scripts/seed-memory-repo.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

scripts/seed-memory-repo.jsView on unpkg

src/core/failure-classifier.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = src/core/failure-classifier.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

src/core/failure-classifier.jsView on unpkg

src/sync/git-sync.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = src/sync/git-sync.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

src/sync/git-sync.jsView on unpkg

scripts/auto-setup.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = scripts/auto-setup.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

scripts/auto-setup.jsView on unpkg

scripts/inject-memory.mjsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = scripts/inject-memory.mjs matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

scripts/inject-memory.mjsView on unpkg

src/core/manifest.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.3.0 matchedPath = src/core/manifest.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.3.0 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.

src/core/manifest.jsView on unpkg

src/core/repl-executor.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = src/core/repl-executor.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

src/core/repl-executor.jsView on unpkg

src/core/worker.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = src/core/worker.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

src/core/worker.jsView on unpkg

scripts/generate-graph.mjsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = scripts/generate-graph.mjs matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

scripts/generate-graph.mjsView on unpkg

src/bench/locomo.jsView file

•matchType = normalized_sha256 matchedPackage = agentic-cortex@7.4.0 matchedPath = src/bench/locomo.js matchedIdentity = npm:YWdlbnRpYy1jb3J0ZXg:7.4.0 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.

src/bench/locomo.jsView on unpkg