project-intelligence-root-cause-analysis
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes content from an external source (the 'Change Events' SharePoint list) to perform causal analysis, scoring, and recommendation generation. Malicious instructions embedded in event descriptions or logs could attempt to influence the agent's logic or output.
- Ingestion points: The skill reads from the 'Change Events' list on the SharePoint site to perform tasks in Steps 1 through 4.
- Boundary markers: Absent. The skill does not instruct the agent to use specific delimiters or to treat the ingested list data as untrusted content that should not be followed as instructions.
- Capability inventory: The skill can update list items in SharePoint (Step 7a), create Word documents in the 'Project Documents' library (Step 7b), and overwrite sections of the 'SHAREPOINT.md' file (Step 7c).
- Sanitization: Absent. No evidence of validation, escaping, or filtering of the external list content was found before the data is processed or persisted back to the environment.
Audit Metadata