The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a cross-session handoff mechanism that automatically ingests data from a local file to restore context. This creates a surface for indirect prompt injection if the handoff file contains instructions meant to override agent behavior.
Ingestion points: The script hooks/session_start.py reads handoff/latest-handoff.md at the beginning of each session.
Boundary markers: No explicit delimiters or boundary markers for the handoff content are defined in the instructions.
Capability inventory: The skill is granted access to Bash, Read, Write, Glob, and Grep tools.
Sanitization: There is no evidence of sanitization or validation of the content read from the handoff files.
[COMMAND_EXECUTION]: The skill relies on local shell scripts and Python hooks for its core functionality and environment configuration.
Evidence: The skill executes tools/mnemonic-paths to resolve the root directory and blackboard paths, and uses automated hooks (hooks/stop.py, hooks/session_start.py) to manage session state.

blackboard