write-slices
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill focuses on local project management tasks, specifically reading and writing within the
spec/features/directory. - [SAFE]: The skill employs security hardening by setting
disable-model-invocation: true, which prevents the agent from autonomously invoking other tools or skills during the decomposition process. - [SAFE]: Command execution is limited to retrieving local git metadata (commit SHA and branch name) to ensure the generated slices are tied to a specific version of the codebase.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided
spec.mdfiles. While it lacks explicit sanitization markers, the risk is mitigated by the highly structured processing logic (parsing specific headers like 'Behaviour' and 'Seams') and the strict instruction to 'implement nothing' and produce only a JSON file as output.
Audit Metadata