echo-ingest-knowledge
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a routing system that coordinates multiple ingestion workflows (newsletter, PRD, research). It does not contain malicious code or dangerous operations.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data such as transcripts, webpages, and social media exports. To mitigate risks, it includes explicit instructions in SKILL.md and references/privacy-boundaries.md to treat this material as evidence rather than instructions, effectively shielding the agent from embedded malicious prompts.
- [COMMAND_EXECUTION]: The skill references a project-specific CLI tool (echo_ingest_knowledge.py) for artifact management and normalization. Usage of this tool is governed by the references/cli-contract.md, which requires explicit operator scope and prohibits the invention of replacement commands.
- [DATA_EXFILTRATION]: Strict privacy guidelines are enforced in references/privacy-boundaries.md. The skill explicitly forbids contacting external parties or publishing data from ingest output without approval, and requires stopping for operator intervention if a workflow needs private sources or outbound actions.
Audit Metadata