The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-generated Markdown files from the local filesystem to manage task data.
Ingestion points: Reads .md task files from the configured tasks_folder (as seen in Workflow 2).
Boundary markers: None explicitly defined for task bodies; the skill prioritizes structured YAML frontmatter for lifecycle operations.
Capability inventory: Filesystem read/write, local network communication via web_fetch, and tool-based task operations.
Sanitization: No explicit sanitization or natural language instruction filtering is applied to the markdown content.
[SAFE]: The skill demonstrates secure credential management by instructing the agent to request API tokens from the user or point to documentation for setup, specifically advising against reading secrets from internal plugin configuration files like data.json.
[SAFE]: Network activity is restricted to the local loopback interface (127.0.0.1) for communicating with the TaskNotes HTTP API, which prevents the exfiltration of sensitive data to external domains.

tasknotes