Skip to content

techne:research-grounded

Flag decisions (technology choices, data release and privacy, study design, access-rule readings, public-claim wording) that lack a # research(YYYY-MM): provenance comment, then web-search to ground them before they harden into code.

Part of techne: /plugin install techne@techne, then /techne:setup.

When to use

  • Before a plan's decisions turn into commits or a release: "are these choices grounded in current best practice or the governing source, or in training-cutoff recall?"
  • After drafting IMPL.md / ROADMAP.md or research-log entries that pick a tool, pattern, protocol, data-release form, exclusion rule, licence reading or public claim.
  • When a decision is stated as fact ("X supports Y") that was never verified against the target's current docs, the class of miss that becomes revertable work.
  • Auditing planning/ design notes, research logs and ADRs for un-cited choices.

Usage

Invoke by name in Claude Code:

/techne:research-grounded

Scope is IMPL.md + ROADMAP.md in the current repo, plus every file, directory or glob in the decision_docs: key of the ## research_grounded block of .claude/skill-context.md (research logs, registered decisions, ADRs, specs). Long append-only logs are audited by recent entries only. Narrow or redirect by naming a file or directory:

/techne:research-grounded planning/

The skill greps for decision language only as a candidate seed, then reads around each hit to keep the genuine choices among alternatives: technology, pattern, architecture, data release, study design, access-rule readings and claim wording. Descriptive prose ("reads X instead of Y") and hypotheticals ("what if we used…") are filtered out. For each committed decision lacking an adjacent research(YYYY-MM): tag or an in-record citation, it reports a decision → gap → action entry and, on confirmation, web-searches current best practice and adds the provenance tag, or no source found when the search finds none. It never invents a citation.

See also

  • techne:docsync: verifies documentation claims against the code; this skill is the sibling that verifies design decisions against evidence.
  • techne:deslop: tightens AI-slop prose, a different kind of doc-quality pass.

Full instructions

SKILL.md on GitHub holds the full instructions Claude follows for this skill.