How do I install the Aidex Decision skill?

Run npx skillmds@latest add yacb2/aidex-decision in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Aidex Decision safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Aidex Decision?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Aidex Decision free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Aidex Decision?

yacb2 (@yacb2) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Aidex Decision by yacb2 — Agent Skill

Name: Aidex Decision
Author: yacb2

Aidex Decision

Use when the user has made an architectural, technical, or product decision and wants it recorded as a written `.context/decisions/` ADR — what was chosen, why, the alternatives, and the consequences — so the team does not re-litigate it later. Fires on "we decided X", "document this decision", "write an ADR", "record the decision to X", "log why we chose X", "we settled on X, write it up". Not for: planning multi-step work (aidex-plan); deferring or parking an idea for later (aidex-backlog); capturing a stakeholder/client request (aidex-request), research notes (aidex-research), or a settled reference (aidex-reference); ecosystem audits (aidex); project-state audits (aidex-audit).

yacb2 a0ecf82 3 files · 10.5 KB Updated Sep 17, 2026

File contents

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-decision" before continuing. Then proceed with the rest of the skill body as normal. If empty/unset, ignore this block.

Decision

Record an architectural, technical, or product decision as an ADR in .context/decisions/ so it is not re-litigated later. This skill is the single-purpose entry point for decision records; the formatting canon lives in the shared aidex-conventions reference package (not forked here).

Workflow

Read the decision conventions canon: ~/.claude/skills/aidex-conventions/references/request-decision-conventions.md (Decisions section; or .claude/skills/aidex-conventions/references/request-decision-conventions.md if a project-level copy exists).
Create a single dated file: .context/decisions/YYYY-MM-DD-<slug>.md (slug kebab-case; date YYYY-MM-DD).
Front-matter per the canon — decisions use the ADR status enum (accepted · superseded · dropped), not the base open/doing/done vocabulary. Set superseded_by: decision/<filename> when one decision replaces another.
Body per the canon template: Context, Decision, Consequences are the mandated triad. Capture the deciding factor, not just the verdict. Add a separate Options Considered (Pros/Cons) section only when two or more alternatives were genuinely weighed; otherwise fold the alternatives into the Decision prose ("considered X and Y; chose Z because …"). Write the artifact in English (canon §Language).

Closing a decision

Decisions close only when superseded or dropped (an accepted ADR stays active in place). Close atomically (stamps updated, sets status + superseded_by, archives per D-10) instead of hand-editing:

bash ~/.claude/skills/aidex-conventions/scripts/close-dated-artifact.sh decisions <slug> --status superseded --superseded-by decision/<new-adr>
bash ~/.claude/skills/aidex-conventions/scripts/close-dated-artifact.sh decisions <slug> --status dropped

Self-check

Validate the artifact you just wrote and fix any violation before closing:

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

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 work	`aidex-plan`
Defer / park / shelve an idea for later	`aidex-backlog`
Capture a stakeholder/client request	`aidex-request`
Investigate / research how something works	`aidex-research`
Document a settled system reference	`aidex-reference`
Audit the Claude Code ecosystem	`aidex`
Audit project state (UX/security/perf/a11y)	`aidex-audit`

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

---name: aidex-decisiondescription: Use when the user has made an architectural, technical, or product decision and wants it recorded as a written `.context/decisions/` ADR — what was chosen, why, the alternatives, and the consequences — so the team does not re-litigate it later. Fires on "we decided X", "document this decision", "write an ADR", "record the decision to X", "log why we chose X", "we settled on X, write it up". Not for: planning multi-step work (aidex-plan); deferring or parking an idea for later (aidex-backlog); capturing a stakeholder/client request (aidex-request), research notes (aidex-research), or a settled reference (aidex-reference); ecosystem audits (aidex); project-state audits (aidex-audit).---> **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-decision"` before continuing. Then proceed with the rest of the skill body as normal. If empty/unset, ignore this block.# DecisionRecord an architectural, technical, or product decision as an ADR in`.context/decisions/` so it is not re-litigated later. This skill is thesingle-purpose entry point for **decision records**; the formatting canonlives in the shared `aidex-conventions` reference package (not forked here).## Workflow1. Read the decision conventions canon: `~/.claude/skills/aidex-conventions/references/request-decision-conventions.md` (Decisions section; or `.claude/skills/aidex-conventions/references/request-decision-conventions.md` if a project-level copy exists).2. Create a single dated file: `.context/decisions/YYYY-MM-DD-<slug>.md` (slug kebab-case; date `YYYY-MM-DD`).3. Front-matter per the canon — decisions use the ADR status enum (`accepted` · `superseded` · `dropped`), **not** the base open/doing/done vocabulary. Set `superseded_by: decision/<filename>` when one decision replaces another.4. Body per the canon template: **Context**, **Decision**, **Consequences** are the mandated triad. Capture the deciding factor, not just the verdict. Add a separate **Options Considered** (Pros/Cons) section only when two or more alternatives were genuinely weighed; otherwise fold the alternatives into the Decision prose ("considered X and Y; chose Z because …"). Write the artifact in English (canon §Language).## Closing a decisionDecisions close only when superseded or dropped (an `accepted` ADR stays activein place). Close atomically (stamps `updated`, sets status + `superseded_by`,archives per D-10) instead of hand-editing:```bashbash ~/.claude/skills/aidex-conventions/scripts/close-dated-artifact.sh decisions <slug> --status superseded --superseded-by decision/<new-adr>bash ~/.claude/skills/aidex-conventions/scripts/close-dated-artifact.sh decisions <slug> --status dropped```## Self-checkValidate the artifact you just wrote and fix any violation before closing:```bashpython3 ~/.claude/skills/aidex-conventions/scripts/validate.py --type decisions```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 work | `aidex-plan` || Defer / park / shelve an idea for later | `aidex-backlog` || Capture a stakeholder/client request | `aidex-request` || Investigate / research how something works | `aidex-research` || Document a settled system reference | `aidex-reference` || Audit the Claude Code ecosystem | `aidex` || Audit project state (UX/security/perf/a11y) | `aidex-audit` |## Related- **aidex-conventions** — owns the shared documentation canon (this skill delegates into its `references/request-decision-conventions.md`).

Frequently asked questions