/doc - Document Findings
SkillDocs & knowledgeDocument findings, research, or knowledge into ProductSpecification/ with proper cross-references. Use when the user wants to write up research results, create a reference doc, consolidate scattered knowledge, or mentions /doc command. Also triggers on "document this", "write this up", "let's capture what we learned", "create a reference for".
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the /doc - Document Findings skill
What this skill tells your AI
The instructions your AI receives, as published by rakovi4/continue-framework in .claude/skills/doc/SKILL.md and read by ahel’s review.
Capture knowledge from a session into a well-structured document under ProductSpecification/, with bidirectional cross-references to related specs, stories, and tasks.
Why This Exists
Research, API discoveries, and architectural findings get lost across conversations. This skill ensures knowledge is captured once, in one canonical location, with links from every document that needs it — so future sessions find the information without re-discovering it.
Usage
/doc # Interactive — asks what to document
/doc "task ordering strategy" # Topic hint
Workflow
Round 1: What and Where
Gather from user:
-
What are we documenting? Summarize the findings in 1-2 sentences. If the current conversation already contains research/findings, extract the summary automatically and confirm with the user.
-
Where should it live? Propose a location based on the topic:
- Existing folder (e.g.,
api-reference/) → add a new.mdfile or append to existing - New folder under
ProductSpecification/→ create{topic-slug}/README.md - Inside a task folder (e.g.,
tasks/35-bug-.../research.md) → already scoped
Decision aid: If the findings are specific to one task/story, keep them in that folder. If they're reusable across multiple stories or tasks, create a shared location.
- Existing folder (e.g.,
-
Filename: Propose based on content (e.g.,
research.md,README.md,{topic}.md). Confirm with user.
Round 2: Cross-References
Identify automatically, confirm with user:
-
What should reference this doc? (inbound links)
- Search for related story
endpoints.mdfiles, taskspec.mdfiles, andCLAUDE.mdthat would benefit from a link to this document. - Use
Grepto find files mentioning the key concepts being documented. - Propose specific files and the line/section where the link should be added.
- Search for related story
-
What should this doc reference? (outbound links)
- Scan the conversation for source materials: story interviews, task specs, external URLs, code files.
- Search
ProductSpecification/for related existing docs. - List proposed references for user confirmation.
Present both lists to the user. They can add/remove entries.
Round 3: Write and Link
-
Write the document with this structure:
- Title and summary
- Findings organized in tables and sections (prefer tables for structured data)
- Links to source materials (outbound references from Round 2)
- External URLs where applicable
-
Update referencing docs — add links in the files identified in Round 2 (inbound references). Use relative paths. Add links in contextually appropriate sections — don't just append to the end.
-
Update indexes:
- If a new folder was created → update
ProductSpecification/CLAUDE.mdstructure tree - If the doc is in a task folder → no index update needed
- If a new folder was created → update
-
Commit all changes with message:
docs: {short description}
Rules
- All documentation in English (project convention)
- Use relative paths for cross-references (
../../tasks/35-bug-.../research.md) - Prefer tables for structured data (endpoints, permissions, field mappings)
- Include external URLs where relevant (API docs, third-party references)
- Don't duplicate content that already exists — link to it instead
- Keep each document focused on one topic; create separate docs for separate topics
Signals
- GitHub stars
- 51
- Forks
- 14
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
doc-rakovi4- Source
- github.com/rakovi4/continue-framework