coding-agent-docs-optimization
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and analyze external documentation pages and machine-readable specifications (e.g., OpenAPI, AsyncAPI). This exposes the agent to indirect prompt injection if the audited documentation contains malicious instructions.
- Ingestion points: Document fetching in SKILL.md (Workflow step 1) and references/agent-surfaces.md (Access before discovery).
- Boundary markers: None. The skill does not instruct the agent to use delimiters or ignore instructions found within the fetched data.
- Capability inventory: The skill uses outbound HTTP requests (likely via curl or wget) and file system writes to devrel-context.md for task management (SKILL.md).
- Sanitization: None. The skill processes fetched content for auditing without sanitizing or escaping potentially executable natural language instructions.
- [COMMAND_EXECUTION]: The skill workflow involves using shell commands to verify that documentation surfaces are accessible to agent crawlers. These commands are a core part of the documentation audit process.
- [DATA_EXFILTRATION]: The skill performs network operations to fetch content from external URLs provided by the user for auditing. While these requests are legitimate and necessary for the skill's function, they represent outbound network activity to non-whitelisted domains.
Audit Metadata