ccf-pipeline-orchestrator
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied artifacts and a status tracking file (
ccfa.yaml) to plan project stages and route tasks to other agents. This creates an attack surface where malicious instructions embedded in project artifacts could influence the orchestration logic or be passed to downstream skills. - Ingestion points: Project state file (
ccfa.yaml), user-provided research artifacts, and project descriptions. - Boundary markers: The skill employs an 'Intake Protocol' and stage classification logic to bound requests, although it lacks explicit delimiters to isolate untrusted data during processing.
- Capability inventory: The skill is capable of reading and suggesting updates to local project files and generating coordination reports and handoff instructions for other agents.
- Sanitization: It implements a privacy boundary classification system (public, private, or mixed) to control data exposure and restricts external search queries based on these levels.
Audit Metadata