applying-coding-baseline
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains explicit instructions to bypass a verification tool named
guarding-agent-directives. While the stated rationale is that the rules are 'pre-vetted', this instruction directs the agent to ignore a security or auditing mechanism designed to vet instructions. - Evidence in SKILL.md: 'Skips guarding-agent-directives verification because the rule set is already pre-vetted.'
- Evidence in SKILL.md: 'This skill does not invoke guarding-agent-directives.'
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted data from project files (
CLAUDE.md,AGENTS.md) to determine where to apply behavioral rules. - Ingestion points: The skill reads
CLAUDE.md,AGENTS.md, and any files referenced or imported by them within the project repository. - Boundary markers: No explicit boundary markers or 'ignore embedded instructions' warnings are used when processing the existing content of these files.
- Capability inventory: The skill has the capability to modify core agent directive files (
CLAUDE.md,AGENTS.md) and create new files. - Sanitization: There is no evidence of sanitization or filtering of the content read from the project files before it is processed by the agent to make replacement decisions.
Audit Metadata