research-technical-docs

Trace technical assertions to version-pinned primary sources and save cited Markdown findings.

1|Updated Jul 13, 2026
One-click install
npx skills add https://github.com/dineshrevunuru/SuperSkills --skill research-technical-docs-dineshrevunuru
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-technical-docs
Source: https://github.com/dineshrevunuru/SuperSkills/tree/main/research-technical-docs
Command: npx skills add https://github.com/dineshrevunuru/SuperSkills --skill research-technical-docs-dineshrevunuru

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Technical answers stated from memory are often wrong, outdated, or version-mismatched, and they fail silently in production months later. This Skill forces every load-bearing technical claim about a library, framework, or API to be traced back to the primary source that owns it — official docs, source code, specs, or a live API call — before you build on it. ## Core Features & Use Cases - Source-of-record ladder: Ranks sources from Tier 1 (running code, versioned official references, specs) down to Tier 4 (Stack Overflow, blogs, other AI answers), which may only be used as leads, never citations. - Trace-gate lint: Every load-bearing assertion must be traced to a Tier 1–2 authority, version-pinned with a permalink, checked against the running artifact where cheap, and flagged under-evidenced if it cannot be traced. - Two modes: A scrappy path for one-fact lookups (API signature, config default, deprecation check) and a rigor path for answers that gate shipped code or architecture decisions. - Use Case: Before forcing a tool call in an Anthropic Messages API router, verify the exact tool_choice object shape against the versioned API reference and one real call, then save the cited finding so the next build reuses it instead of re-researching. ## Quick Start Ask the agent to verify how a specific API or library feature behaves in your pinned version and save the answer as a cited Markdown file.

Frequently Asked Questions about research-technical-docs

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify an API signature or config default before using it?

Pin the exact version from your lockfile, then trace the fact to the versioned official reference or the source code, and confirm with one real API call where cheap. Record the assertion with a permalinked citation and the date checked.

What sources should I cite when researching technical documentation?

Cite Tier 1 sources first: source code, actual API responses, specs, and versioned official references. Tier 2 covers official guides and changelogs; maintainer issue comments are Tier 3. Stack Overflow, blogs, and other AI answers are leads only, never citations.

What should I do when official docs disagree with the source code?

Trust the running artifact over the document — docs lag code. Record the disagreement as a finding, cite the code or live behavior with a pinned commit, and note the doc as stale rather than silently picking one side.

When should I use the scrappy path versus the rigor path?

Use scrappy for a one-fact lookup nothing downstream depends on, like a single enum value. Use rigor when the answer gates code you will ship or an architecture decision; when unsure, choose rigor because untraced facts fail silently later.

When is this skill the wrong tool for a research question?

It only covers external technical authorities like docs, specs, and dependency source code. Market or competitor questions belong to business research, user behavior questions belong to UX research methods, and understanding your own codebase belongs to codebase analysis or debugging.