skill-test
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The skill uses the shell command pattern ('!') to execute a local bash script ('yaml-helper.sh') at load time to resolve project configuration. This allows for code execution before the agent processes the markdown body.
- [COMMAND_EXECUTION]: The skill executes multiple local scripts ('agent-check.sh', 'compat-inventory.sh', 'yaml-helper.sh') via the Bash tool to perform its checks. These scripts are located within the project's internal '.claude' and 'hooks' directories.
- [INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface as it ingests and evaluates the instructions of other skills and agents. • Ingestion points: Reads all '.claude/skills//SKILL.md' and '.claude/agents/.md' files. • Boundary markers: None specified to separate the analyzed skill instructions from the tester's own logic. • Capability inventory: Includes 'Bash', 'Write', and 'Edit' tools. • Sanitization: No explicit sanitization or escaping of the ingested content is mentioned.
Audit Metadata