Agent skill
tech-impl-doc
Write precise technical implementation docs (developer or user) with clear structure, accurate flows, render-safe Mermaid diagrams, and audience-appropriate detail; avoid boilerplate; include a meaningful Code Index for developer docs.
Install this agent skill to your Project
npx add-skill https://github.com/acking-you/myclaude-skills/tree/main/skills/tech-impl-doc
SKILL.md
Technical Implementation Doc Skill
When to use
- The user asks for an implementation doc, technical deep dive, architecture/design write-up, or a user guide describing how a system works.
Confirm up front
- Audience: user / developer / mixed.
- Scope: which features or subsystems to cover.
- Output location and language.
- Diagrams: include Mermaid for multi-step flows unless the user says not to.
Audience rules
User-facing docs
- No code locations or internal class names unless explicitly requested.
- Focus on configuration, behavior, and operational steps.
- Use direct clickable links (raw URLs). Do not hide references behind footnotes.
- Avoid boilerplate: no “this document describes…”, no “version info”, no filler headings.
Developer-facing docs
- Include code paths and line numbers.
- Add a Code Index: each entry must explain what the code is responsible for.
- Call out constraints, failure modes, and trade-offs.
Structure rules
- Headings must be meaningful (no parenthetical fluff like “(quick copy)” or “(notes)”).
- Prefer numbered sections for complex systems.
- Keep intro minimal; skip generic explanations.
Mermaid rules (avoid parse errors)
- Keep node labels simple; avoid parentheses and heavy punctuation in labels.
- If a label must contain special characters, wrap it in quotes.
- Avoid line breaks inside labels.
- Use simple node IDs (A, B, C…).
Safe example:
flowchart TB
A["buildRemotePlan (query_port)"] --> B["CrossClusterMergeStep"]
Workflow
- Identify audience and constraints.
- Scan existing docs; consolidate instead of duplicating.
- Verify claims against source code or authoritative docs.
- Draft core flow first (request path → execution → merge → output).
- Add Mermaid diagrams for key flows.
- Add constraints/unsupported features section.
- Add Code Index with descriptions (developer docs only).
- Final pass to remove boilerplate and meaningless headings.
Output checklist
- Audience-appropriate detail level.
- No redundant intro/version sections unless requested.
- Mermaid diagrams render without parse errors.
- Code Index entries explain their purpose (developer docs only).
- Links are clickable in user docs.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
github-wrapped
Generate a verifiable GitHub Wrapped year-in-review as a single-file HTML (raw gh API JSON saved, Python-built dataset embedded), with a Bilibili-style narrative, smooth transitions, and mobile-friendly paged mode. Requires gh CLI to be authenticated.
codex-session-history
Search Codex local session history stored in SQLite and session_index.jsonl. Use this skill when the user wants to list, search, filter, or inspect locally saved Codex sessions by session id, provider, time range, preview text, thread name, or archived status.
gemini-image
Generate images using AI when user wants to create pictures, draw, paint, or generate artwork. Supports text-to-image and image-to-image generation.
excalidraw
Generate hand-drawn style diagrams (architecture, flowcharts, system design) as .excalidraw.json files. Use when user wants diagrams, mentions Excalidraw, or needs Mermaid-to-visual conversion.
research
Structured web research with multi-source validation. Output reports with complete citation links. Focus on key results, verify information quality.
gen-commit-msg
Generate concise commit messages based on conversation context and minimal git inspection.
Didn't find tool you were looking for?