The Agent Skills Directory

[SAFE]: The skill demonstrates defensive engineering practices by explicitly instructing the agent to treat external reports and data as "evidence, not instructions" in references/issue-intake.md, which helps prevent indirect prompt injection attacks.
[SAFE]: Instructions for diagnosing failures include explicit requirements to redact secrets and sensitive payloads before sharing evidence across process or service boundaries (references/diagnosis-probes.md).
[SAFE]: All external documentation links refer to official, well-known services and organizations such as AWS, Microsoft, SQLite, and the Hypothesis project, which are recognized as trusted sources.
[SAFE]: The skill identifies vendor-specific storage paths (e.g., .qp/alaga/intake/) consistent with the expected author context for the developer 'quantipixels'.
[SAFE]: The use of Yoruba nomenclature for system roles (e.g., Alága, Atọ́nà, Àtúnwò) is linguistically consistent with the skill's branding and does not constitute a homoglyph attack or obfuscation technique.

alaga