capture-tasks-from-meeting-notes
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from meeting notes to generate Jira tasks, which is the primary intended functionality.\n
- Ingestion points: Untrusted data enters the agent context through Confluence page content (fetched via
getConfluenceContent) or directly pasted text (Step 1).\n - Boundary markers: The workflow establishes a clear verification boundary in Step 5, requiring the user to confirm the parsed items before the agent proceeds.\n
- Capability inventory: The skill utilizes
createJiraIssueto perform write operations in the Jira environment.\n - Sanitization: The skill mitigates the risk of processing malicious instructions by mandating a human review step (Step 5: 'Wait for Confirmation') where the user must explicitly approve, skip, or modify each action item before execution.
Audit Metadata