The Agent Skills Directory

[SAFE]: The skill defines a limited set of commands (list, show, quote) for interacting with a local CLI tool (recall). No unauthorized command execution, network exfiltration, or sensitive file access was detected. The command usage follows the intended purpose of browsing a knowledge repository.
[SAFE]: The skill ingests data from external entries (curated knowledge captures), which constitutes an indirect prompt injection surface. However, given the primary purpose of the skill is knowledge retrieval and the absence of high-risk capabilities like writing to files or making network requests based on that content, the risk is assessed as safe.
Ingestion points: Entries are retrieved and displayed to the agent using the recall loa show <id> and recall loa quote <id> commands described in SKILL.md.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for the ingested content to prevent instructions within the entries from being followed.
Capability inventory: The skill is restricted to reading and displaying content from the local Library of Alexandria database.
Sanitization: No sanitization or validation of the retrieved content is mentioned in the skill instructions.

source-command-recall-loa