daily-briefing
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external web searches and internal memory files to generate a briefing, creating a surface for indirect instructions to influence the agent.
- Ingestion points: The skill reads from memory/pending-tasks.md, memory/active-threads.md, memory/user-profile.md, and gathers data via the WebSearch tool (SKILL.md).
- Boundary markers: No specific delimiters or "ignore instructions" markers are used to isolate the data being summarized from the skill's operational instructions.
- Capability inventory: The skill is limited to Read, Glob, WebSearch, and WebFetch tools (SKILL.md). The absence of file-write or command-execution tools significantly mitigates the risk of an injection causing local system harm.
- Sanitization: There is no evidence of sanitization or filtering of the content retrieved from internal files or the web before it is processed by the model.
Audit Metadata