setup-matt-pocock-skills

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands including git remote -v, gh (GitHub CLI), and glab (GitLab CLI) to perform repository discovery and issue management operations.
  • [DATA_EXFILTRATION]: The skill reads repository configuration files such as .git/config and command outputs to identify remote repository URLs. While intended for local configuration, this behavior involves accessing and processing sensitive repository metadata.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from the repository (e.g., git remote URLs, existing documentation content) and incorporates this information into its decision-making process and generated output without explicit boundary markers or sanitization.
  • Ingestion points: Reads git remote -v, .git/config, AGENTS.md, CLAUDE.md, and directory structures (docs/adr/, .scratch/) in SKILL.md Step 1.
  • Boundary markers: Absent; the instructions do not specify delimiters or warnings to ignore malicious instructions embedded in the read files.
  • Capability inventory: File system write access (CLAUDE.md, AGENTS.md, docs/agents/*.md) and shell execution (git, gh, glab).
  • Sanitization: Absent; the skill summarizes found repository data directly into the user dialogue and documentation files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 10:16 AM
Security Audit — agent-trust-hub — setup-matt-pocock-skills