The Agent Skills Directory

[SAFE]: The skill implements strong security practices by providing dedicated scripts for redacting sensitive data. redact_handoff.py and project_session_events.py use comprehensive regex patterns to identify and mask API keys, tokens, private keys, and local user paths.
[COMMAND_EXECUTION]: The skill utilizes local Python scripts for processing session data. These scripts rely exclusively on the Python standard library and do not engage in network communication or perform unsafe system operations. scripts/smoke_test_handoff.py uses importlib.util for internal script testing, which is a standard and safe practice.
[PROMPT_INJECTION]: The skill ingests data from local logs and transcripts, which presents a surface for indirect prompt injection.
Ingestion points: scripts/project_session_events.py processes content from files provided by the user.
Boundary markers: Not present; the generated prompt does not use explicit delimiters to encapsulate the potentially untrusted summarized content.
Capability inventory: The skill's capabilities are limited to text processing and stdout output, lacking network access or file system modification tools.
Sanitization: The skill performs automated redaction of credentials and local paths using predefined patterns in scripts/redact_handoff.py.

session-handoff-prompt