retro
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes session transcripts and subagent logs, which are untrusted data sources. Malicious instructions could be embedded in these logs to influence the agent's reporting phase. Ingestion points: Session transcript files, IDs, and databases (SKILL.md, Rule 1). Boundary markers: The skill lacks explicit instructions to treat transcript content as untrusted text, though it instructs the agent to condense the data first (Rule 2). Capability inventory: Reading file system/database, generating reports, and executing scripts for visualization. Sanitization: None explicitly defined beyond extraction of specific fields (time, duration, tool input/error).
- [DYNAMIC_EXECUTION]: The instructions require the agent to generate and run a script to create a text-based chart for subagent activity, which involves runtime code generation and execution. Evidence: 'Render the chart with a script from the condensed timeline; never place bars by hand.' (SKILL.md, Rule 9).
Audit Metadata