specstory-session-summary

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use standard command-line utilities including ls, grep, head, and tail to discover and read session history files located in the .specstory/history/ directory.
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes session history files which store logs of previous interactions. These logs represent an attack surface for indirect prompt injection if a prior session processed untrusted content from the web or external repositories that contained hidden instructions.
  • Ingestion points: The skill reads file content from the .specstory/history/*.md path as seen in Step 1 and Step 2 of the instructions.
  • Boundary markers: The instructions do not provide specific delimiters or "ignore instructions" warnings to the agent when processing the content of these history files.
  • Capability inventory: The skill uses tools for listing and reading files (ls, grep, head, tail). It does not demonstrate capabilities for network exfiltration (e.g., curl) or writing/modifying system files.
  • Sanitization: There is no evidence of content validation, escaping, or filtering applied to the text extracted from the history files before the agent processes it for the summary.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 08:14 AM
Security Audit — agent-trust-hub — specstory-session-summary