adr
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes the current codebase, implementation details, and git history to automatically populate ADR content. This process is susceptible to indirect prompt injection if malicious instructions are present in the files or commit messages being read.\n
- Ingestion points: The skill explicitly reads the codebase, file implementation details, and
git logoutput to infer architectural context (SKILL.md, Step 4).\n - Boundary markers: No explicit delimiters or isolation instructions are provided to distinguish between the data being processed and the skill's operational logic.\n
- Capability inventory: The skill can execute local shell commands (
git log) and write to the filesystem (SKILL.md, Step 6).\n - Sanitization: The skill mitigates risk by requiring the agent to present the full generated document to the user for review and confirmation before writing any files to disk.\n- [COMMAND_EXECUTION]: The skill is instructed to execute the
git logcommand to extract historical context for the decisions. This represents a dependency on local environment command execution.
Audit Metadata