Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/thimslugga-adr
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/adr
What's inside
- Skill name
adr- Description
- Create and maintain Architecture Decision Records (ADRs). Use when the user proposes a significant architectural change, records a refactoring decision, resolves a technical debate, or asks to write, update, deprecate, or supersede an ADR.
- License
- MIT
- Compatibility
- Requires python.
- Instructions
- 111 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (4)
- SKILL.md
- adr-template.md · assets/
- index-template.md · assets/
- validate_adrs.py · scripts/
SKILL.md (first lines)
# Architecture Decision Records (ADR) Skill
Write and maintain ADRs: lightweight plain-text files capturing architectural
choices, their context, and their consequences.
## Why ADRs exist
Architecture consists of the decisions that are hard to change. ADRs address
two failures:
- **Context loss.** Future developers cannot see why code was written a certain
way, and remove constraints they do not know exist. ADRs record the rationale.
- **Discussion decay.** ADRs preserve the conclusion and the forces that shaped
it, without replaying the debate.
## Rules
1. **Keep it to one or two pages.** An ADR records a decision. It is not a
design specification.
2. **Treat the body as immutable.** Once an ADR is committed, its prose stands
as the record of what was decided at that time. Only the `status`,
`supersedes`, and `superseded_by` fields may change afterward.
3. **State the downsides.** Every choice has costs. List negative consequences,
technical debt incurred, and new constraints accepted.
## Creating an ADR
1. Determine the number from ADR filenames in `docs/adr/`. Increment the highest
four-digit prefix, or start at `0001` when none exist. The frontmatter `id`,
H1 heading, and filename prefix must carry the same number.
2. Name the file `NNNN-short-descriptive-title.md` and write it to `docs/adr/`.
Use lowercase kebab case after the numeric prefix.
3. Build the body from `assets/adr-template.md`. Keep every heading at its
template level and in order. Where a section does not apply, write one
sentence explaining why rather than removing or leaving the heading empty.
4. Set `status`:
- `proposed` — under review, not yet agreed
- `accepted` — agreed and in force
- `rejected` — considered and declined; keep the record because the
reasoning has value
5. Add a row to the index. Rewrite only the content between the
`BEGIN GENERATED INDEX` and `END GENERATED INDEX` markers in
`docs/adr/INDEX.md`, preserving everything outside them. Include exactly one
row per ADR, sorted by ascending numeric ID, with values copied verbatim from
its filename and frontmatter. Create the index from `assets/index-template.md`
if it does not exist.
## Superseding a decision
To reverse or replace an accepted decision, make three edits:
1. Write a new ADR with the next number, setting `supersedes: [NNNN]`.
2. On the old ADR, set `status: superseded` and `superseded_by: [NNNN]`.
Change nothing else in that file.
3. Update both rows in the index.
Use `status: deprecated` instead when a decision is abandoned with no
replacement, leaving `superseded_by` empty.
## Validate the ADR set
Run the bundled validator from this skill directory, passing the project's ADR
directory:
```shell
python scripts/validate_adrs.py <project-root>/docs/adr
```
Use `--json` for machine-readable output. Exit code `0` means the ADR documents,
supersession links, and generated index agree; exit code `1` means validation
failed. Fix every reported error before finishing.
The validator checks deterministic structure and cross-document consistency. It
does not judge page length, prose quality, factual tone, or body immutability;
review those separately because they require human judgment or version history.
## Where documents live
```text
docs/
├── adr/
│ ├── INDEX.md # index, one line per ADR
│ ├── 0001-use-structured-logging.md
│ └── 0002-migrate-to-sqlite-cache.md
├── rfc/
│ ├── INDEX.md # index, one line per RFC
│ └── 0001-plugin-permission-model.md
└── design/
└── type-enrichment-pipeline.md
```
## Before finalizing
- [ ] The forces and constraints that drove the decision are stated.
- [ ] Discarded alternatives are listed with the reason each lost.
- [ ] The Negative consequences list is populated with specifics.
- [ ] The tone is factual, with no unquantifiable praise ("elegant", "clean", "perfect").
- [ ] The index row is added and its number matches the file.
- [ ] `scripts/validate_adrs.py` reports `VALID`.
## Resources
- **`assets/adr-template.md`** — skeleton for ADR files
- **`assets/index-template.md`** — skeleton for `docs/adr/INDEX.md`
- **`scripts/validate_adrs.py`** — deterministic ADR and index validatorFrom thimslugga/agent-plugins, MIT.
More from thimslugga/agent-plugins
- 0.567score
golang-pro
thimslugga
Write, edit, update, or review Go code to a consistent professional standard: gofumpt-formatted, go vet clean, golangci-lint clean under a curated config.
SkillAFreshness: AClaude Code +9 - 0.565score
gauntlet-loop
thimslugga
Turns any goal into one short, paste-ready "gauntlet loop" prompt - a prompt that makes an agent set a concrete quality bar, split the work into small…
SkillAFreshness: AClaude Code +9 - 0.565score
ast-grep-outline
thimslugga
Use when exploring or modifying a codebase and you need a cheap structural map of files, directories, imports, exports, or direct members before reading full…
SkillAFreshness: AClaude Code +9 - 0.540score
impeccable
thimslugga
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or…
SkillAFreshness: AClaude Code +9 - 0.529score
redshift-api
thimslugga
Run SQL against Amazon Redshift — submit statements, poll status, page through results, and browse databases/schemas/tables.
SkillBFreshness: BClaude Code +9 - 0.529score
ai-firstify
thimslugga
Analyze, re-engineer, or bootstrap projects to align with AI-first design principles.
SkillBFreshness: BClaude Code +9 - 0.529score
migrate-agent-plugin
thimslugga
Migrate an existing Claude Code, Codex, Cursor, Kiro, Copilot, VS Code, or other client-specific agent plugin to the portable Agent Plugins v1 structure while…
SkillBFreshness: BClaude Code +9 - 0.524score
economist-style
thimslugga
Apply The Economist style guide to written content.
SkillBFreshness: BClaude Code +9 - 0.524score
autoresearch
thimslugga
'Autonomous iterative experimentation loop for any programming task.
SkillBFreshness: BClaude Code +9 - 0.524score
grill-me
thimslugga
Structured questioning to reach shared understanding before action.
SkillBFreshness: BClaude Code +9 - 0.524score
handoff
thimslugga
Compacts the current conversation into a handoff document so a fresh agent can continue the work in a new session.
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 14, 2026 (14 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/thimslugga/agent-plugins
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from thimslugga
- 0.567score
golang-pro
thimslugga
Write, edit, update, or review Go code to a consistent professional standard: gofumpt-formatted, go vet clean, golangci-lint clean under a curated config.
SkillAFreshness: AClaude Code +9 - 0.565score
ast-grep-outline
thimslugga
Use when exploring or modifying a codebase and you need a cheap structural map of files, directories, imports, exports, or direct members before reading full…
SkillAFreshness: AClaude Code +9 - 0.565score
gauntlet-loop
thimslugga
Turns any goal into one short, paste-ready "gauntlet loop" prompt - a prompt that makes an agent set a concrete quality bar, split the work into small…
SkillAFreshness: AClaude Code +9 - 0.540score
impeccable
thimslugga
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or…
SkillAFreshness: AClaude Code +9 - 0.529score
redshift-api
thimslugga
Run SQL against Amazon Redshift — submit statements, poll status, page through results, and browse databases/schemas/tables.
SkillBFreshness: BClaude Code +9 - 0.529score
aws-iam
thimslugga
Verified corrections for IAM behaviors that AI agents frequently get wrong — policy evaluation edge cases, trust policy gotchas, STS session limits…
SkillBFreshness: BClaude Code +9
Cite this page
adr skill. Agent Mart. https://agentmart.dev/skills/thimslugga-adr (last updated 2026-10-24)