The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of an external Python package, claude-session-index, which is used to index and search session history.
[INDIRECT_PROMPT_INJECTION]: The skill reads and summarizes historical session logs, which creates a potential surface for indirect prompt injection.
Ingestion points: The agent uses the sessions context command as described in SKILL.md to retrieve logs from past sessions for analysis.
Boundary markers: No specific delimiters or instructions are provided to the agent to distinguish between the content of the logs being analyzed and the agent's own task instructions.
Capability inventory: Across all files, the skill utilizes Bash, Read, Grep, and AskUserQuestion tools.
Sanitization: The skill does not implement any filtering or sanitization of the retrieved session logs before they are processed by the agent.

session-recap