page-qa
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads landing page assets from the workspace for evaluation, creating a surface where malicious content embedded in the analyzed files could attempt to influence the agent's assessment. 1. Ingestion points: workspace/pages/[page-name]/strategy.json, copy.md, index.html, meta.json, and visuals/index.md. 2. Boundary markers: Absent; the skill does not instruct the agent to ignore or delimit instructions found within the data files. 3. Capability inventory: The skill writes to qa.md and executes a local shell script (SKILL.md, eval-summary.sh). 4. Sanitization: Absent; the skill does not specify validation or filtering for the data it ingests.
- [COMMAND_EXECUTION]: The skill includes a local bash script, scripts/eval-summary.sh, used for parsing metadata and generating summaries via jq and grep. The script uses unvalidated positional arguments to construct file paths, which presents a minor risk of path traversal within the workspace if used with untrusted parameters.
Audit Metadata