lode-session-recap

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes raw conversation history to generate structured log entries.
  • Ingestion points: The agent reads the full conversation history for the current session to extract decision context and technical changes (SKILL.md).
  • Boundary markers: The instructions lack explicit boundary markers or directions for the agent to ignore or delimit instructions found within the conversation data.
  • Capability inventory: The skill possesses file-write capabilities through the execution of the scripts/lode_raw.py helper script (scripts/lode_raw.py).
  • Sanitization: There is no evidence of sanitization or validation of the ingested conversation content before it is used to populate log entries.
  • [COMMAND_EXECUTION]: The skill uses shell commands to execute a local Python script (scripts/lode_raw.py) for resolving configuration paths and appending session data to JSON files. This is a core part of its intended logging functionality.
  • [EXTERNAL_DOWNLOADS]: The script scripts/lode_raw.py includes an optional dependency on the yaml (PyYAML) package for parsing configuration files.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 09:55 AM
Security Audit — agent-trust-hub — lode-session-recap