doc-fix_skill
- Shell
0
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill shotaiuchi/dotclaude --skill doc-fix- SKILL.md11.8 KB
Overview
This skill fixes issues identified in docs/reviews/*.md review files and applies the suggested changes to the original documents. It supports interactive single-file fixes and parallel batch processing using a doc-fixer sub-agent for multiple files. The tool updates review tables with a Status column so changes are traceable.
How this skill works
The skill locates review files (docs/reviews/*.md) or derives review paths from provided original paths, parses High/Medium priority tables and Future Considerations lists, and maps each review to the original file (md, yaml, yml, json, txt). For single-file interactive runs it presents detected issues for selection and applies chosen fixes directly. For --all or multiple-file runs it launches doc-fixer sub-agents (up to 5 concurrent) to apply all suggested fixes and then records results in the review files.
When to use it
- You have a docs/reviews/*.md file containing structured review comments and suggestions.
- You want to apply all suggested fixes to many review files quickly using parallel processing.
- You prefer to pick specific high- or medium-priority items to fix interactively for a single file.
- You need the review files updated with fix status for traceability and audit.
- You need automated mapping from review filenames to original document paths (supports common extensions).
Best practices
- Run without arguments to auto-discover docs/reviews/*.md and review detected files before applying fixes.
- Use --all when processing multiple files or when you want non-interactive, consistent fixes via sub-agents.
- For single-file changes, use interactive mode to confirm ambiguous suggestions before applying.
- Keep review tables in the specified template format so the parser reliably extracts issues.
- If many files exist, process in batches to avoid exceeding MAX_PARALLEL (5) and to review partial results.
Example use cases
- Apply all fixes across the entire reviews directory: /doc-fix --all
- Fix a single review interactively and choose which items to apply: /doc-fix docs/reviews/commands.wf0-status.md
- Process a subset using globbing: /doc-fix docs/reviews/*.md --all
- Retry failed files after a batch run: /doc-fix docs/reviews/docs.API.md docs/reviews/docs.CONFIG.md --all
- Generate traceable updates by marking fixed items with Fixed (YYYY-MM-DD) in the review tables.
FAQ
The tool reports an error for that review file. In parallel mode it skips or marks the task failed so other files continue processing.
How many sub-agents run in parallel?
Up to MAX_PARALLEL (5) sub-agents run concurrently. Larger sets are processed in batches.
Does the tool change the review files?
Yes. After applying fixes it adds a Status column and marks fixed rows with Fixed (YYYY-MM-DD) for traceability.