p-activity-digest
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from external communication platforms.
- Ingestion points: Untrusted data enters the agent context through Slack search results and channel reads (Step 2) and Fireflies meeting transcripts and summaries (Step 3) as described in
SKILL.md. - Boundary markers: Absent. The synthesis logic in Step 4 of
SKILL.mddoes not specify the use of delimiters or instructions to the model to ignore potential commands embedded within the fetched data. - Capability inventory: The skill is granted the
save to filesystempermission in theSKILL.mdfrontmatter and utilizes file-writing tools in the output phase (Step 5). - Sanitization: Absent. There is no evidence of content filtering, escaping, or validation performed on the messages or transcripts before they are processed by the LLM.
Audit Metadata