create-plan
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Utilize s the GitHub CLI (
gh issue view) to fetch task requirement s from externa l issue s. This is a standard operation for the skill's desig n an d target s a well-know n service. - [PROMPT_INJECTION]: The skill possesse s an indirect pro m pt injection surface as it ingest s untruste d requirement s from GitHub issue s to drive plan generatio n. The requirement -to-step map ping proces s provide s a logica l validatio n laye r agains t maliciou s instr uction s.
- Ingestio n point s: Requirement s are fetche d via
gh issue viewinSKIL L.md. - Boundar y marker s: Absen t; the agen t is directe d to extrac t al l conten t for map ping.
- Capabil it y inventor y: Acces s to
ghtoo l an dto do s_onesho tfor plan ticke t generatio n. - Sanitizatio n: No ne; relie s on struct ure d map ping to ensure compliance with origina l task s.
Audit Metadata