gsd-add-todo
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted text from conversation history and user arguments to generate file content and Git commits.
- Ingestion points: Captured text from {{GSD_ARGS}} and conversation context in the add-todo process.
- Boundary markers: None; the skill lacks delimiters or instructions to treat input as data rather than instructions.
- Capability inventory: The skill performs file creation, directory manipulation, and git commit operations.
- Sanitization: No sanitization, escaping, or validation logic is specified for the ingested content.
- [COMMAND_EXECUTION]: The skill uses tools to perform filesystem changes (creating directories and files) and version control actions (git commits) based on instructions defined in an external workflow file.
- [REMOTE_CODE_EXECUTION]: The skill's operational logic is defined in an external file at C:/Users/nicol/.codex/get-shit-done/workflows/add-todo.md, making the skill's behavior dependent on the contents of that local file rather than the skill package itself.
Audit Metadata