agents-md-craft
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by reading and restructuring existing project documentation files, which constitutes a data ingestion surface.
- Ingestion points: The workflow involves inspecting target instruction files (AGENTS.md, CLAUDE.md) and project layout facts.
- Capability inventory: The skill has the ability to write to the file system and execute shell commands to verify documentation accuracy.
- Sanitization: Instructions emphasize maintaining authorization boundaries and distinguishing between documenting a command and executing it, though no automated sanitization is specified.
- [COMMAND_EXECUTION]: The skill allows for the execution of repository-specific commands to validate documentation content.
- Evidence: SKILL.md directs the agent to run commands when needed to resolve material uncertainty.
- Evidence: checklist.md advises running commands for verification only when authorized.
- Context: This capability is restricted to auditing documentation and includes prose warnings against unauthorized operations like deployments or migrations.
Audit Metadata