qcsd-cicd-swarm
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external pipeline artifacts to perform quality analysis and flag detection. \n- Ingestion points: PIPELINE_ARTIFACTS directory (referenced in SKILL.md and steps/01-flag-detection.md). \n- Boundary markers: None identified; the instructions do not specify the use of delimiters or 'ignore' instructions for the ingested artifacts. \n- Capability inventory: Spawns sub-agents via the Task tool (Steps 2, 4, 8) and executes shell commands via the aqe CLI tool (Steps 1, 7). \n- Sanitization: No specific sanitization or validation of content within the pipeline artifacts is defined. \n- [COMMAND_EXECUTION]: The skill employs the aqe CLI utility for memory retrieval and storage tasks. \n- Evidence: 'aqe memory search' in steps/01-flag-detection.md and 'aqe memory store' in steps/07-learning-persistence.md. \n- Context: These commands are used for legitimate state persistence and data retrieval within the Agentic QE framework, which is consistent with the skill's primary purpose.
Audit Metadata