The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs extensive network operations to retrieve research materials from the internet. It uses tools such as WebSearch and mcp__web-reader__webReader to query external domains and fetch full-page content.
[COMMAND_EXECUTION]: The skill interacts with the local file system by reading configuration data from .loom/config.yaml and writing synthesized markdown reports to the current directory or specified vault paths.
[PROMPT_INJECTION]: Indirect Prompt Injection surface detected. The skill's primary function involves ingesting and analyzing untrusted content from the public web (defined in Phase 1: Scout). Malicious instructions embedded in these external sources could potentially influence the agent's analysis or output in later phases.
Ingestion points: Web content crawled via WebSearch and webReader as described in Workflows/01-scout.md.
Boundary markers: Absent; the instructions do not provide delimiters or warnings for the LLM to ignore instructions found within retrieved web data.
Capability inventory: The agent has permission to Read local files and Write results to disk as defined in SKILL.md.
Sanitization: The workflow does not include specific content validation or escaping logic for the retrieved external data before it is processed.

survey