context-surfing

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill implements a SessionStart hook that executes a local shell script (scripts/handoff-checker.sh) to detect and alert the agent about pending handoff files at the start of a session.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes various local project artifacts to serve as a 'wave anchor,' interpolating their content into the agent's context.
  • Ingestion points: Reads .context-surfing/handoff-*.md, AGENTS.md, README.md, and session plan artifacts.
  • Boundary markers: The hook script wraps findings in <context-surfing-handoff> XML-like tags, though the handoff files themselves rely on standard markdown headers.
  • Capability inventory: The skill interacts with CLI tools such as gh, npx, and entire for installation and session state management.
  • Sanitization: Content from project artifacts and session handoffs is loaded without specific sanitization or escaping mechanisms.
  • [EXTERNAL_DOWNLOADS]: The skill's documentation and installation procedures reference GitHub repositories (pskoett/pskoett-skills) and the entire CLI tool (entireio/cli) for session persistence.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 08:41 AM
Security Audit — agent-trust-hub — context-surfing