create-local-plan
Pass
Audited by Gen Agent Trust Hub on Mar 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git logandgit diffto analyze the local repository's history and current state. These commands are standard for research and are constrained to the local file system. - [PROMPT_INJECTION]: The skill processes untrusted content from the codebase and git logs, creating a surface for indirect prompt injection.
- Ingestion points: Project files and git history (SKILL.md).
- Boundary markers: Absent; the agent lacks instructions to ignore malicious directives found within analyzed content.
- Capability inventory: Writing plan files to the
.github/plans/directory (SKILL.md). - Sanitization: None; ingested content is not validated or escaped before being processed into new documentation.
Audit Metadata