setup-matt-pocock-skills

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill reads repository configuration from .git/config and git remote -v to identify the environment context.
  • [COMMAND_EXECUTION]: Explicitly instructs the agent to use CLI tools like gh and glab to interact with external issue trackers.
  • [PERSISTENCE]: Modifies existing agent configuration files (CLAUDE.md or AGENTS.md) to embed instructions for other skills, ensuring settings persist across sessions.
  • [INDIRECT_PROMPT_INJECTION]: Creates configuration files that influence the behavior of other agents, representing an indirect injection surface.
  • Ingestion points: Reads repository and agent configuration from files like .git/config and CLAUDE.md (SKILL.md).
  • Boundary markers: The generated markdown files lack explicit delimiters or instructions to ignore embedded content (issue-tracker-github.md, issue-tracker-gitlab.md, domain.md).
  • Capability inventory: Performs file writes and executes shell commands via established CLI tools (SKILL.md).
  • Sanitization: Employs a workflow requiring user confirmation before performing final file writes (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 11:18 AM
Security Audit — agent-trust-hub — setup-matt-pocock-skills