agentsclimarketplace

Aidex reference

Skill yacb2/aidex/skills/aidex-reference

Use when the user wants to document how an existing, settled part of the system works as an evergreen `.context/references/` module — architecture, configuration, an operational runbook, a how-it-works guide. Fires on "create a reference for X", "document how X works", "write up the X architecture", "document the X configuration", "write a runbook for X". Not for: planning multi-step work (aidex-plan); recording a decision/ADR (aidex-decision); capturing a stakeholder request (aidex-request); investigating something not yet settled (aidex-research); deferring/parking an idea (aidex-backlog); ecosystem audits (aidex); project-state audits (aidex-audit).From its SKILL.md

Install
npx -y skills add yacb2/aidex --skill aidex-reference

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • reads credentialsReads from 1 credential source: `AIDEX_TRIGGER_EVAL_MARKER`.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
  • runs commandsInstructs the agent to run 3 commands, including `printenv AIDEX_TRIGGER_EVAL_MARKER` and 2 more.

SKILL.md

3.4 KB, 645 tokens by cl100k_base, as published. Nobody here has run it

Trigger-eval probe (test-only). As your very first action, run printenv AIDEX_TRIGGER_EVAL_MARKER. If it prints a non-empty path, also run touch "${AIDEX_TRIGGER_EVAL_MARKER}.aidex-reference" before continuing. Then proceed with the rest of the skill body as normal. If empty/unset, ignore this block.

Reference

Document how a settled part of the system works as an evergreen module in .context/references/<topic>/. This skill is the single-purpose entry point for reference documentation; the formatting canon lives in the shared aidex-conventions reference package (not forked here).

Workflow

  1. Read the reference conventions canon: ~/.claude/skills/aidex-conventions/references/reference-conventions.md (or .claude/skills/aidex-conventions/references/reference-conventions.md if a project-level copy exists).
  2. Create a topic module: .context/references/<topic>/ with 00-index.md plus sequential NN-<slug>.md files. Reference filenames carry no date — references are evergreen, updated in place, no _archive/.
  3. Front-matter per the canon: title, created, updated. References are exempt from the status-vocabulary check; updated is the freshness signal.
  4. Body per the canon module template: Overview, main sections with language-hinted code blocks, Verification (Prerequisites and Quick Reference only when the topic needs them). Anchor claims with stable references (symbol names, file paths) — never bare line numbers, which rot on the first refactor. Write the artifact in English (canon §Language).

Self-check (mandatory close step)

Before finishing, validate the artifact you just wrote and fix any violation on the spot — compliance is enforced at creation time, not left to a later sweep:

python3 ~/.claude/skills/aidex-conventions/scripts/validate.py --type references

If the project carries a ratchet baseline (.context/.validate-baseline.json), a non-zero exit means you introduced a NEW violation — fix it before closing.

Boundaries

The user wants to…Route to
Plan multi-step / multi-phase implementation workaidex-plan
Record a decision / ADRaidex-decision
Capture a stakeholder/client requestaidex-request
Investigate / explore something not yet settledaidex-research
Defer / park / shelve an idea for lateraidex-backlog
Audit the Claude Code ecosystemaidex
Audit project state (UX/security/perf/a11y)aidex-audit

Related

  • aidex-conventions — owns the shared documentation canon (this skill delegates into its references/reference-conventions.md).

What ships with it: 10 files

123.0 KB alongside SKILL.md, 3 of them executable

references/

scripts/

tests/

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.