!
Purpose & Capability
The skill name/description say it normalizes whitespace and trims text, but main.py does not implement any text processing: it only prints a JSON object echoing the provided --input and --mode. The declared outputs (cleaned_text) are not produced. This is a direct mismatch between claimed capability and actual code.
!
Instruction Scope
SKILL.md instructs running 'python main.py --input sample.txt --mode clean' implying the tool will read/process a file or input text. main.py treats --input as a string or path but does not read files or perform cleaning. The SKILL.md also contains a templated risk statement (looks like an unresolved template) and a minor formatting artifact in the usage example ('\b ash'), indicating sloppy packaging. Instructions therefore over-promise relative to runtime behavior.
✓
Install Mechanism
No install spec and no third-party dependencies; the skill is instruction-only with a single small Python file using only argparse and json. There are no downloads or install steps that add risk.
✓
Credentials
The skill requests no environment variables, no credentials, and no config paths. There is no evidence of requests for unrelated secrets or excessive permissions.
✓
Persistence & Privilege
The skill does not request always:true, does not modify other skills or system settings, and has no install-time persistence. Autonomous invocation defaults are unchanged (normal).
What to consider before installing
This package appears to be a stub/placeholder rather than a working text-cleaner. If you need a text-normalization skill, do not rely on this in production until the author provides code that actually performs the described transformations. Before installing or invoking: (1) request the full source or a link to a repository showing the implementation, (2) verify main.py reads input and outputs cleaned_text (not just metadata), (3) confirm version consistency (registry metadata says 1.0.0 but skill.json lists 0.1.0) and fix unresolved template text in SKILL.md, and (4) run the code in a safe, isolated environment to validate behavior. The current footprint is low-risk but functionally misleading.