aif-archive
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform system-level operations including moving directories (mv), creating archive structures (mkdir), and querying the git log (git) to determine completion dates. These commands are restricted to the intended archival workflow.\n- [INDIRECT_PROMPT_INJECTION]: The skill evaluates plan files and roadmaps to detect completed tasks, creating a surface for indirect instructions to influence the agent.\n
- Ingestion points: The skill reads file content from any markdown plan located in paths.plans/ and the project's ROADMAP.md.\n
- Boundary markers: There are no explicit delimiters or specific instructions provided to the agent to ignore or isolate potential commands or adversarial text embedded within the tasks or phase descriptions.\n
- Capability inventory: The skill possesses significant capabilities including the ability to write/edit files and execute scoped Bash commands, which could be leveraged if the agent follows instructions found in data files.\n
- Sanitization: The skill performs simple regex checks for task completion (e.g., matching '- [x]') but does not sanitize or filter the remaining content of the files it processes.
Audit Metadata