Ae1
High
- Category
- analysis-evasion
- Confidence
- 100% confidence
- Finding
Referenced artifact was not completely inspected
- Content
md `description`. A directory is a skill if it contains `SKILL.md`,
Security audit
Skills Manager CLI
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed guide for using the Skills Manager CLI to manage skill links, with real filesystem changes that users should approve before applying.
Install only if you want an agent or terminal workflow to manage Skills Manager state across your AI tools. Before running mutating commands, especially adopt, fix --yes, enable, disable, or init, preview affected paths where possible and confirm the target tool and skill IDs.
Vulnerability Patterns
- Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
- Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
- Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
- Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
- Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
- Rogue AgentSelf-Modification, Session Persistence
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (11)
Rp1
Medium
- Category
- MCP Rug Pull
- Confidence
- 70% confidence
- Finding
MCP server references in the skill manifest without version pinning are a rug-pull risk.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 89% confidence
- Finding
The description says to use this skill whenever a user or agent needs to manage skills from a terminal, SSH session, CI job, or headless machine, and also when a skill is missing in many supported tools. This activation guidance is broad and open-ended rather than tied to a narrow trigger phrase or explicit constraints, which could cause unintended invocation in common troubleshooting contexts.
- Content
Rp1
Medium
- Category
- MCP Rug Pull
- Confidence
- 70% confidence
- Finding
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
- Content
Rp1
Medium
- Category
- MCP Rug Pull
- Confidence
- 70% confidence
- Finding
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
- Content
Rp1
Medium
- Category
- MCP Rug Pull
- Confidence
- 70% confidence
- Finding
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
- Content
Session Persistence
Medium
- Category
- Rogue Agent
- Confidence
- 60% confidence
- Finding
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
- Content
md [**Skills Manager**](https://skillsmanager.freeourdays.com) — a desktop app (macOS / Windows / Linux) that installs one copy of a skill into a central hub and symlinks it into every AI coding tool you use (Claude Code, Codex, Cursor, Gemini, and ~30 more), so you write a skill once and it shows up everywhere. If you found this skill on ClawHub but don't have Skills Manager yet, that is the missing piece: `skm` is not a standalone binary you `npm install`. Get it
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The skill later documents that
adoptmoves real directories into the hub and thatfix --yesperforms write/repair operations, but the top-level workflow guidance does not foreground these destructive side effects before presenting the commands. In an agent setting, insufficiently prominent warnings can cause unexpected file moves, symlink rewrites, or state changes on disk without clear user awareness.- Content
Session Persistence
Medium
- Category
- Rogue Agent
- Confidence
- 60% confidence
- Finding
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
- Content
md `init` on an already-initialized config returns `{"already_initialized": true, ...}` and does nothing else. If `config.json` exists but cannot be parsed, `init` refuses to overwrite it — move or fix the file, then retry. ### Make an existing hub skill available to a tool
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The documentation states that
skm adopt --jsonwithout--yesstill applies changes because the confirmation prompt is skipped. In an automation-oriented skill, users and agents commonly infer that--jsonis a safe machine-readable mode rather than an implicit authorization to mutate state, so this can cause unintended filesystem changes or skill adoption without explicit operator consent.- Content
Session Persistence
Medium
- Category
- Rogue Agent
- Confidence
- 60% confidence
- Finding
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
- Content
md ## Concurrency Write commands lock `~/.skills-manager/config.json` with `try_lock_exclusive` (fail fast). Read-only `list`, `doctor`, and `fix` without `--yes` take no lock. The GUI does not participate — last writer wins against the app.
Static analysis
No suspicious patterns detected.