The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts, specifically extract_today.py and obsidian-helper.py, using the uv run command. These scripts are bundled with the skill and perform necessary tasks such as parsing local session logs and managing the Obsidian diary integration.
[DATA_EXFILTRATION]: The skill accesses local agent session files in ~/.pi and ~/.omp and transmits metadata/summaries via the nmem synchronization tool to aggregate activity across multiple machines. This behavior is consistent with the skill's primary stated purpose of cross-machine session aggregation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes conversational data from agent sessions which could potentially contain untrusted instructions.
1. Ingestion points: Reads local .jsonl session files and remote thread content retrieved via the nmem_read_thread tool.
1. Boundary markers: The instructions do not provide explicit delimiters or specific warnings for the agent to ignore embedded instructions within the ingested session content.
1. Capability inventory: The skill has the capability to write files to the Obsidian vault and execute local Python scripts.
1. Sanitization: No explicit sanitization, escaping, or filtering of the ingested session text is implemented before it is processed for the daily summary.

daily-recap