feature-intake

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from stakeholder requirements and existing GitHub issues to generate new issue content. 1. Ingestion points: Stakeholder requests (user input), remote GitHub issue searches (SKILL.md Step 3), and local project documentation (SKILL.md Step 4). 2. Boundary markers: None identified; untrusted data is directly interpolated into drafting templates. 3. Capability inventory: The skill can create GitHub issues (SKILL.md Step 9) and modify GitHub Project boards (SKILL.md Step 9). 4. Sanitization: No input validation or output escaping is performed on the ingested content before it is used to draft issue bodies.
  • [COMMAND_EXECUTION]: The skill relies on executing shell commands to interact with the system environment and GitHub services. 1. Evidence: Uses the GitHub CLI (gh) for repository and project management, Git for checking branch state, and ripgrep (rg) for searching local files. 2. Mitigation: The skill correctly requires explicit user confirmation before performing any write operations to the repository or project board, which helps mitigate the risk of unauthorized actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 11:19 AM
Security Audit — agent-trust-hub — feature-intake