setup-matt-pocock-skills
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from the repository's file system and external issue trackers, creating a potential surface for indirect prompt injection if those sources contain malicious instructions.
- Ingestion points: Repository metadata files (CLAUDE.md, AGENTS.md, GLOSSARY.md), git configuration files, and data retrieved via GitHub (gh) or GitLab (glab) CLI tools (such as issue titles and bodies).
- Capability inventory: Execution of shell commands via git, gh, and glab; file system writes for configuration and documentation files; and label creation via tracker APIs.
- Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands in the ingested data.
- Sanitization: The skill provides proactive mitigation in issue-tracker-gitlab.md, instructing the agent to use structured JSON via file APIs rather than interpolating data into shell commands. It also recommends using safe file APIs or quoted heredocs for multi-line GitHub issue bodies.
- [COMMAND_EXECUTION]: The skill utilizes legitimate development tools (git, gh, glab) to perform its setup tasks. The use of these tools is scoped to the skill's stated purpose of repository configuration and follows established conventions for these command-line interfaces.
Audit Metadata