plan-feature-from-issue

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data from GitHub issue titles, bodies, and comments to generate product specifications. An attacker could potentially embed malicious instructions within an issue to influence the agent's planning output.
  • Ingestion points: GitHub issue data is retrieved via the gh issue view command as described in Step 0.
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when processing the issue body.
  • Capability inventory: The skill has the capability to write to the filesystem (SPEC files), modify roadmap documents, and call other agent skills like design-feature.
  • Sanitization: No explicit sanitization or filtering of issue content is mentioned before it is used to fill rubric slots.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute the GitHub CLI (gh issue view) to retrieve project information. This is an expected and legitimate developer tool usage for the skill's stated purpose.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves data from GitHub, which is a well-known and trusted service. This network interaction is necessary for the skill to function and follows standard development workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 11:43 AM
Security Audit — agent-trust-hub — plan-feature-from-issue