planning-summarization

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes Bash to execute curl commands for interacting with ServiceNow REST API endpoints.- [DATA_EXFILTRATION]: As part of its intended functionality, the skill accesses and retrieves sensitive business data including portfolio budgets, project costs, and contract details from a ServiceNow instance to produce management reports.- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface due to the ingestion of user-controlled data from ServiceNow records into the agent's summarization context.
  • Ingestion points: Untrusted data enters the context through fields such as short_description, description, and acceptance_criteria from various ServiceNow tables like dm_demand and rm_story.
  • Boundary markers: The summarization templates in Step 7 and Example 1 lack explicit delimiters or instructions to treat external data as untrusted content.
  • Capability inventory: The skill utilizes MCP tools and Bash for data retrieval and status aggregation.
  • Sanitization: There is no evidence of sanitization or filtering applied to the retrieved database content before it is interpolated into the summarization prompt.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 02:24 PM
Security Audit — agent-trust-hub — planning-summarization