The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill includes a hardcoded authentication secret in the configuration file.
Evidence: The .env file contains the line MEMOS_ACCESS_TOKEN=memos_pat_KgThSDbQeBJorb8OX7LA7QadbRjc6kOv.
[COMMAND_EXECUTION]: The skill requires the agent to execute local JavaScript scripts via a runtime to interact with the API.
Evidence: docs/setup.md and SKILL.md instruct the agent to execute $RUNTIME "$API_SCRIPT" to perform all memos-related actions.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing untrusted data from a remote Memos instance without adequate protection.
Ingestion points: Memo content and comments are retrieved from an external API in scripts/actions/memo.cjs and scripts/actions/comment.cjs.
Boundary markers: Absent. Retrieved memo content is displayed directly to the agent's context without delimiters or specific instructions to ignore embedded commands within the data.
Capability inventory: The skill allows the agent to execute shell commands (running the api.cjs script) and perform authenticated network operations (CRUD operations on the Memos instance).
Sanitization: While scripts/sanitize.cjs provides keyword-based redaction for secrets (e.g., tokens, passwords), it does not sanitize content to prevent structural prompt injection or command sequence escapes.

memos