The Agent Skills Directory

[PROMPT_INJECTION]: The skill fetches data from an external Memos instance which creates a surface for indirect prompt injection if the fetched content contains malicious instructions.
Ingestion points: scripts/memos.sh fetches memo content and comments via curl from the user-configured MEMOS_URL.
Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions when presenting memo content to the agent.
Capability inventory: The script can read, create, update, and delete memos and reactions, and uses curl, jq, and python3 for processing.
Sanitization: While jq is used for JSON structure handling, the actual text content is not sanitized to prevent instruction injection.
[COMMAND_EXECUTION]: The skill relies on executing local system commands to perform its operations.
Evidence: scripts/memos.sh executes curl, jq, and python3 to interact with the REST API.
Context: The script uses python3 -c snippets for URL encoding and JSON formatting, which are standard utility patterns in shell-based skills.

memos