post-incident-review
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from incident reports, forensic findings, and communication logs, creating a potential surface for indirect prompt injection.
- Ingestion points: External incident response reports, chronologies, forensic analysis results, and communication logs as described in the 'Context the Agent Needs' section and Step 2.
- Boundary markers: Section 8 ('Prompt Injection Safety Notice') provides instructional guardrails directing the agent to treat data as informational rather than executable, though it does not mandate specific technical delimiters (e.g., XML tags or unique string wrappers) for the processed data.
- Capability inventory: The skill uses
Read,Grep, andGlobtools to access and process the data. - Sanitization: The skill relies on natural language instructions to prevent the agent from following directives embedded within the analyzed incident data.
Audit Metadata