architecturehardparts
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a local Python script, 'scripts/newadr.py', which the agent is instructed to execute to generate Architecture Decision Record (ADR) files.
- [PROMPT_INJECTION]: The skill's workflow involves reading repository evidence (code, schemas, logs) for analysis, creating an indirect prompt injection surface. 1. Ingestion points: 'SKILL.md' workflow steps 2 and 3 involve reading repository modules, schemas, and manifests. 2. Boundary markers: Absent. 3. Capability inventory: Shell execution of the internal 'scripts/newadr.py' script and read/write access to the local filesystem. 4. Sanitization: The helper script uses safe literal lambda replacements in its regex substitution logic to prevent injection.
Audit Metadata