The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to evaluate and triage external content including code diffs, research evidence, and writing samples. This creates a surface where the agent could ingest malicious instructions embedded within the data being analyzed.
Ingestion points: Content provided for triage in the change, research, and writing categories within SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are defined in the skill instructions.
Capability inventory: The skill has access to the exec tool, allowing it to run shell commands.
Sanitization: No sanitization or validation logic is specified for the ingested content.
[COMMAND_EXECUTION]: The skill uses exec to run vendor-specific CLI tools like algal-skills and algal via the bunx package runner for checking capabilities and managing authentication.
[EXTERNAL_DOWNLOADS]: The use of bunx for algal-skills and algal involves fetching and running packages from the registry. These resources appear to be part of the author's own infrastructure and tools intended for the skill's operation.

algal-triage