The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Clojure source files automatically via a trigger, creating a surface where malicious content inside those files could influence agent behavior.
Ingestion points: The workflow in workflows/fix-on-save.yml ingests any file matching the pattern **/*.{clj,cljs,cljc} upon save.
Boundary markers: There are no protective boundary markers or instructions to the agent to ignore potentially malicious embedded content within the source files being processed.
Capability inventory: The skill utilizes tools to read from and write to the file system (e.g., fix_clojure_delimiters) and performs syntax validation.
Sanitization: No evidence of content sanitization or validation of the input source code is present before it is passed to the editing tools.
[METADATA_POISONING]: The SKILL.md file contains misleading information about the skill's capabilities.
Evidence: The documentation explicitly lists fix_clojure_delimiters and validate_clojure_syntax as "(Deprecated/Removed)" tools, yet these exact tools are the primary components defined and executed in the workflows/fix-on-save.yml file. This inconsistency can lead to a misjudgment of the skill's actual operational surface.

clojure-quality