idea-discovery-robot
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements an automated orchestration pipeline that processes data from potentially untrusted sources (user input and external AI responses) and uses it to drive subsequent agent actions.
- Ingestion points: The skill ingests user input through the
$ARGUMENTSvariable and incorporates external critical feedback from a reviewer model (gpt-6-astra) in Phase 5. - Boundary markers: The instructions do not define explicit boundary markers or delimiters when interpolating external data into the prompts for sub-skills such as
/research-litor/idea-creator. - Capability inventory: The skill is capable of triggering multiple automated sub-skills, writing report files to the filesystem (
idea-stage/IDEA_REPORT.md), and managing a project manifest (MANIFEST.md). TheAUTO_PROCEED=trueconfiguration allows the pipeline to progress through these capabilities without requiring user confirmation, which could facilitate the execution of instructions embedded in the processed data. - Sanitization: No sanitization, escaping, or validation of the ingested external content is performed before it is passed to subsequent stages of the pipeline.
Audit Metadata