The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external, potentially untrusted data including repository instructions, design documents, source code entrypoints, and state machine specifications. This creates an attack surface where an adversary could embed malicious instructions within code comments or documentation to influence the agent's behavior during the audit process.
Ingestion points: As described in SKILL.md, the agent reads repository instructions, design/intent documents, as-built specifications, and source code files to build a behavior inventory.
Boundary markers: The instructions do not specify the use of clear delimiters or 'ignore embedded instructions' prompts when processing these external files.
Capability inventory: The skill itself does not define or invoke dangerous tools such as network exfiltration, arbitrary file writing, or shell command execution; its primary function is to generate a CoverageGapReport markdown file.
Sanitization: There is no evidence of sanitization, filtering, or validation of the content read from external project files before it is processed by the agent.

auditing-gameplay-implementation-coverage