slack-digest
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from Slack channels via semantic search and thread fetching tools, which is then processed by the LLM for classification and summarization.
- Ingestion points: The skill uses
slack semantic searchandfetch-slack-threadtools to retrieve message content from external Slack channels (SKILL.md, Workflow Step 1). - Boundary markers: There are no explicit instructions to the agent to treat Slack message content as untrusted data or to ignore embedded instructions found within threads.
- Capability inventory: The skill possesses the capability to generate markdown reports and deliver content as HTML, potentially via S3 uploads (SKILL.md, Workflow Step 5, 6, and Self-Review Checklist).
- Sanitization: The instructions do not specify any sanitization, escaping, or filtering of the Slack thread content before it is interpolated into the final report or analyzed for classification.
Audit Metadata