PR Design Doc
OpenHands/OpenHands
For a non-trivial pull request, write a self-contained HTML design doc under the temporary .pr/ directory and link a visibility-appropriate preview in the PR description, so maintainers grasp the…
Agent skill
by cosmicstack-labs in cosmicstack-labs/mercury-agent-skills
ADR methodology, templates, decision capture workflows, and architectural governance patterns
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-records --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/categories/development/architecture-decision-records .claude/skills/architecture-decision-records && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .claude/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-recordsType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-records --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/categories/development/architecture-decision-records .agents/skills/architecture-decision-records && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .agents/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-records --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/categories/development/architecture-decision-records .cursor/skills/architecture-decision-records && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .cursor/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/cosmicstack-labs/mercury-agent-skills.git --path categories/development/architecture-decision-records--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-records --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/categories/development/architecture-decision-records .gemini/skills/architecture-decision-records && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .gemini/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-recordsInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/categories/development/architecture-decision-records .github/skills/architecture-decision-records && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .github/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install cosmicstack-labs/mercury-agent-skills architecture-decision-records --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/cosmicstack-labs/mercury-agent-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/categories/development/architecture-decision-records .opencode/skills/architecture-decision-records && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "architecture-decision-records" agent skill from https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/development/architecture-decision-records into .opencode/skills/architecture-decision-records/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "architecture-decision-records", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
architecture-decision-recordsADR methodology, templates, decision capture workflows, and architectural governance patterns
Architecture Decision Records is an agent skill from cosmicstack-labs/mercury-agent-skills. ADR methodology, templates, decision capture workflows, and architectural governance patterns
Its SKILL.md is about 3.8k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Development, covering Architecture decision records. The repository describes itself as: A curated registry of reusable Mercury Agent, Open Claw or Hermes Agent skills designed for real developer workflows, persistent memory, and token-efficient execution. The licence is MIT.
4 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 30392fb. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
npxbrewnpmFrom the folder's file list and the shell code blocks in SKILL.md.
Hosts in commands or code, which the agent is likely to contact:
github.comFrom URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Architecture Decision Records loads about 3.8k tokens when it runs. Until then it costs about 31 tokens; SKILL.md has 796 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from cosmicstack-labs/mercury-agent-skills at commit 30392fb, republished under its MIT licence (© cosmicstack-labs). 796 words, ~3,841 tokens.
.claude/skills/architecture-decision-records/SKILL.md (or your agent's skills folder).Capture architectural decisions systematically so your team understands not just what was decided, but why — and what alternatives were considered.
A diagram shows the current architecture. An ADR explains why it is that way. When someone asks "why did we do it this way?" the ADR is the answer.
Every architectural decision exists in a web of constraints, tradeoffs, and alternatives. If you only record the conclusion, future engineers will wonder if you considered the obvious alternative — and they might reverse it without understanding why the original choice was made.
An ADR doesn't need to be a 10-page document. A structured 1-page record is infinitely better than nothing. If the process is heavy, people won't follow it.
Architecture evolves. An ADR that gets superseded is a success — it means the system adapted. Old ADRs remain valuable as historical records of the team's thinking.
| Level | Capture | Storage | Review | Enforcement |
|---|---|---|---|---|
| 1: Tribal | Decisions in Slack/meetings | Nobody remembers | None | None |
| 2: Documented | Some decisions written down | Shared drive or wiki | Sporadic | None |
| 3: Systematic | All significant decisions as ADRs | In repository alongside code | PR review requires ADR for arch changes | Basic: "needs ADR" check |
| 4: Integrated | ADRs linked to implementation | Searchable, indexed, cross-referenced | Mandatory ADR review for arch changes | Automated: lint checks for ADR format |
| 5: Governance | ADRs drive architecture reviews | Catalog with status dashboard | Regular architecture review board | Automated compliance checks |
Target: Level 3 for most teams. Level 4+ for regulated or long-lived systems.
# ADR-{NNN}: {Title}
## Status
[Proposed | Accepted | Deprecated | Superseded]
*If Superseded, list the replacing ADR: Superseded by ADR-{NNN}*
## Context
{Describe the problem, constraints, and forces at play.
What is the business or technical need?
What are the non-negotiable constraints?
What options were considered?}
## Decision
{State the decision clearly.
What are we doing? What are we NOT doing?}
## Consequences
{List the positive and negative consequences of this decision.
What tradeoffs are we accepting?
What becomes easier? What becomes harder?}
## Alternatives Considered
{List alternatives and why they were rejected. This is the most
important section for future readers.}
### Option A: {Name}
- **Pros**: ...
- **Cons**: ...
- **Why rejected**: ...
### Option B: {Name}
- **Pros**: ...
- **Cons**: ...
- **Why rejected**: ...
## Compliance
{How will we verify this decision is followed?
Automated checks? Manual review? Linting rules?}For quick decisions that still need recording:
# ADR-042: Use PostgreSQL for Analytics Store
**Status**: Accepted
**Date**: 2024-03-15
**Author**: Alice Chen
**Deciders**: Alice Chen, Bob Smith, Carol Davis
## Context
We need a store for aggregated analytics data. Requirements:
JSON support, time-series optimized, managed service preferred.
## Decision
Use PostgreSQL with TimescaleDB extension on RDS.
## Rationale
- JSONB for flexible event schemas
- TimescaleDB hypertables for time-series queries
- RDS for managed operations
- Team already familiar with PostgreSQL
## Alternatives
- **MongoDB**: Better for unstructured data, but adds operational complexity
and team lacks expertise → rejected
- **ClickHouse**: Excellent for analytics but overkill for our volume (100k events/day) → rejected
## Consequences
+ Existing PostgreSQL expertise applies
+ Single database reduces operational burden
- Need to learn TimescaleDB syntax
- JSONB queries are less performant than dedicated document store┌────────────┐ ┌──────────────┐ ┌────────────┐ ┌───────────────┐
│ Identify │ │ Draft │ │ Review │ │ Accept & │
│ Decision ─┼─► │ ADR ─┼─► │ & Discuss│──►│ Commit │
│ Needed │ │ (Proposed) │ │ │ │ (Accepted) │
└────────────┘ └──────────────┘ └────────────┘ └───────────────┘
│ │
│ Rejected │ Later
▼ ▼
┌──────────┐ ┌──────────────┐
│ Revise │ │ Superseded │
│ or File │ │ by New ADR │
└──────────┘ └──────────────┘Write an ADR when the decision:
Examples of ADR-worthy decisions:
Examples of non-ADR decisions:
# Create the ADR file
mkdir -p docs/adr/
cp templates/adr-template.md docs/adr/ADR-043-use-graphql-for-public-api.md
# ADR naming convention
# ADR-{NNN}-{short-descriptive-slug}.md
# Use leading zeros for sorting: ADR-001, ADR-002, ..., ADR-043Include the ADR in the same PR as the implementation, or as a standalone PR for purely architectural decisions. Reviewers should check:
# After acceptance, the ADR status changes to "Accepted"
# If the decision is later revisited:
## Status
Superseded by ADR-052
## Rationale for Deprecation
In 2024, a managed Kafka service became available that eliminates
the operational overhead that motivated our original SQS choice.
The scale of our event processing has also grown 10x since ADR-021.project/
├── docs/
│ └── adr/
│ ├── index.md # Catalog of all ADRs
│ ├── ADR-001-initial-project-structure.md
│ ├── ADR-002-database-selection.md
│ ├── ADR-003-api-protocol.md
│ ├── ADR-004-deprecated-by-008.md
│ ├── ...
│ └── ADR-052-event-stream-architecture.md
└── .adr-dir # Points to the ADR directoryWhy store ADRs in the repository:
# Architecture Decision Records
## Active (Accepted)
| ADR | Title | Date | Area |
|-----|-------|------|------|
| ADR-003 | API Protocol: GraphQL | 2024-01-20 | API |
| ADR-002 | Database: PostgreSQL | 2024-01-15 | Data |
| ADR-008 | Event Bus: RabbitMQ | 2024-02-10 | Infrastructure |
## Proposed
| ADR | Title | Date | Author |
|-----|-------|------|--------|
| ADR-009 | Cache Strategy: Redis with write-through | 2024-03-01 | Alice |
## Deprecated / Superseded
| ADR | Title | Superseded By | Date |
|-----|-------|---------------|------|
| ADR-001 | Initial: SQLite | ADR-002 | 2024-01-15 |
| ADR-004 | Event Bus: SQS | ADR-008 | 2024-02-10 |Sometimes one PR involves several related decisions. Handle with care:
# ADR-030: Order Service Decomposition
**Status**: Accepted
## This ADR covers three decisions:
1. Extract order management from the monolith
2. Use event-driven communication between order and inventory services
3. Adopt PostgreSQL for the order service database
## Decision
Extract the Order Service as a standalone service...Alternative: Write one ADR per decision and reference them:
ADR-031: Extract Order Service from Monolith
ADR-032: Event-Driven Communication for Order Service
ADR-033: Database Selection for Order ServiceA concise format for decisions with clear tradeoffs:
## Decision (Y-Statement)
In the context of {situation/need},
facing {constraint/force},
we decided for {option A} over {option B}
to achieve {positive consequence},
accepting {negative consequence}.
---
**Example:**
In the context of needing real-time notifications across services,
facing the constraint of not wanting to manage a dedicated messaging infrastructure,
we decided for AWS SNS over RabbitMQ
to achieve zero operational overhead for pub/sub messaging,
accepting vendor lock-in to AWS and higher per-message costs at scale.Sometimes the most valuable ADR is the one about a decision you didn't take:
# ADR-017: Rejected — Migrate to Microservices
**Status**: Rejected
**Date**: 2024-02-01
## Context
Proposal to break the monolith into microservices for better scalability.
## Decision
We decided NOT to pursue microservice decomposition at this time.
## Rationale
- Team size (6 engineers) is too small to manage N services
- Current monolith handles 10k RPM comfortably
- Deployment frequency is satisfactory (daily)
- Distributed transactions would add complexity without clear benefit
- We'll revisit this when:
a) Team grows to 15+
b) Monolith deployment takes >30 minutes
c) Two or more features need different scaling policies# Architecture Review Board Charter
## Purpose
Ensure architectural consistency and quality across all products.
## Composition
- 1 Staff Engineer (permanent)
- 2 Senior Engineers (rotating, 6-month term)
- 1 Product Manager (non-voting)
## When to Escalate
- Cross-team architectural decisions
- Technology stack additions
- Major refactoring or migrations
- Decisions with significant cost implications
## Process
1. Author drafts ADR → send to ARB
2. ARB reviews within 1 week
3. ARB meeting to discuss (if needed)
4. Decision documented in ADR status# .adr-lint.yml
rules:
required-sections:
- Status
- Context
- Decision
- Consequences
- Alternatives Considered
status-values:
allowed:
- Proposed
- Accepted
- Deprecated
- Superseded
- Rejected
naming:
pattern: '^ADR-\d{3}-[a-z0-9-]+\.md$'
message: "ADR files must follow ADR-{NNN}-{slug}.md naming"
no-duplicate-numbers: true
index-required: true
index-path: 'docs/adr/index.md'# Run ADR linting in CI
npx adr-lint docs/adr/
# Example output:
# ✓ ADR-001: All required sections present
# ✓ ADR-002: All required sections present
# ✗ ADR-003: Missing "Alternatives Considered" section
# ✓ ADR-004: Valid status "Accepted"
# ✗ ADR-005: Invalid naming — use ADR-005-{slug}.md# In code comments, reference the ADR that explains the design choice
# Uses Redis-backed rate limiting (see ADR-022)
# Rationale: We need distributed rate limiting across 10 instances
# and in-memory approaches won't work with horizontal scaling.
from ratelimit import RateLimiter
# SQLite for local dev, PostgreSQL in production (see ADR-002)
if config.ENV == "production":
db = PostgresDatabase(config.DATABASE_URL)
else:
db = SQLiteDatabase(":memory:")
# Using UUID v4 instead of auto-increment IDs (see ADR-015)
# Rationale: Prevents ID enumeration and simplifies sharding
order_id = uuid.uuid4()# ADR-010: Authentication Architecture
## Status
Accepted (Updated 2024-03-01)
## Changelog
| Date | Change | Author |
|------|--------|--------|
| 2024-01-15 | Initial draft | Alice |
| 2024-01-20 | Added SSO requirement | Bob |
| 2024-02-01 | Accepted after ARB review | Carol |
| 2024-03-01 | Updated token expiry from 1h to 24h based on UX feedback | Alice |# adr-tools (command-line)
# https://github.com/npryce/adr-tools
# Install
brew install adr-tools
# Create a new ADR
adr new Use PostgreSQL for analytics store
# Creates: doc/adr/0001-use-postgresql-for-analytics-store.md
# List all ADRs
adr list
# Mark as superseded
adr supersede 0001 0008 # ADR-001 is superseded by ADR-008
# Link ADRs
adr link 0001 "Amends" 0003# Log4brains (modern ADR manager with UI)
# https://github.com/thomvaill/log4brains
# Install
npm install -g @log4brains/cli
# Initialize
log4brains init
# Create ADR
log4brains adr:new
# Preview the knowledge base
log4brains preview
# Build static site
log4brains build© cosmicstack-labs, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in categories/development/architecture-decision-records of cosmicstack-labs/mercury-agent-skills.
Open the folder on GitHubat commit 30392fb
Architecture Decision Records next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Architecture Decision Records this skillcosmicstack-labs/mercury-agent-skills | 476 | — | ~3.8k | Automated safety check: Pass | MIT | |
| PR Design DocOpenHands/OpenHands | 90k | — | ~2.4k | Automated safety check: Pass | MIT | |
| Cto AdvisorIbrahim-3d/orchestrator-supaconductor | 380 | 4 repos | ~2.4k | Automated safety check: Pass | MIT | |
| Improve Codebase Architectureywwynm/EverythingDone | 144 | 15 repos | ~1.3k | Automated safety check: Pass | GPL-3.0 | |
| Domain Modelingbrim-borium/spotify_sdk | 166 | 5 repos | ~806 | Automated safety check: Pass | Apache-2.0 | |
| Design Doc MermaidSpillwaveSolutions/design-doc-mermaid | 175 | 1 repos | ~5.6k | Automated safety check: Pass | None |
OpenHands/OpenHands
For a non-trivial pull request, write a self-contained HTML design doc under the temporary .pr/ directory and link a visibility-appropriate preview in the PR description, so maintainers grasp the…
Ibrahim-3d/orchestrator-supaconductor
Technical leadership guidance for engineering teams, architecture decisions, and technology strategy.
ywwynm/EverythingDone
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/.
brim-borium/spotify_sdk
Build and sharpen a project's domain model. An agent skill from brim-borium/spotify_sdk.
SpillwaveSolutions/design-doc-mermaid
Create Mermaid diagrams (flowchart, sequence, class, ER, state, C4, architecture) from text or source code.
DrCatHicks/learning-opportunities
Facilitates deliberate skill development during AI-assisted coding.
cosmicstack-labs/mercury-agent-skills
Use this before implementing a product, feature, SaaS, AI app, or side project to score product risk and choose the smallest validation step.
cosmicstack-labs/mercury-agent-skills
HyperFrames CLI dev loop — project scaffolding, validation (lint/inspect), browser preview with live reload, MP4/WebM rendering, and environment troubleshooting (doctor, browser, info, upgrade).
cosmicstack-labs/mercury-agent-skills
Asset preprocessing for HyperFrames compositions — local text-to-speech narration (Kokoro-82M, no API key), audio/video transcription (Whisper), and background removal for transparent overlays…
cosmicstack-labs/mercury-agent-skills
Design and implement agent-to-agent handoff protocols for multi-agent systems.
cosmicstack-labs/mercury-agent-skills
Monitor AI agent health, detect anomalies, set up alerting, and maintain observability dashboards for production multi-agent systems.
cosmicstack-labs/mercury-agent-skills
Design and operate task delegation systems for multi-agent fleets.
Categories
ADR methodology, templates, decision capture workflows, and architectural governance patterns. Architecture Decision Records is an agent skill from cosmicstack-labs/mercury-agent-skills.
Architecture Decision Records fits situations like: tasks that involve Architecture decision records.
Run `npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a claude-code`. Or copy the skill folder (categories/development/architecture-decision-records in cosmicstack-labs/mercury-agent-skills) into .claude/skills/architecture-decision-records in your project. Claude Code loads it when a task matches its description.
Run `npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a codex`. Or copy the skill folder (categories/development/architecture-decision-records in cosmicstack-labs/mercury-agent-skills) into .agents/skills/architecture-decision-records in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add cosmicstack-labs/mercury-agent-skills --skill architecture-decision-records -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/architecture-decision-records, .gemini/skills/architecture-decision-records, .github/skills/architecture-decision-records and .opencode/skills/architecture-decision-records in your project.
Going by SKILL.md and its folder, Architecture Decision Records needs the command-line tools its instructions call (npx, brew and npm).
SKILL.md names 1 domain. In commands or code: github.com; the agent is likely to contact it when it follows the instructions. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Architecture Decision Records is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.8k tokens (SKILL.md is roughly 15k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Architecture Decision Records: PR Design Doc (OpenHands/OpenHands, 90k stars), Cto Advisor (Ibrahim-3d/orchestrator-supaconductor, 380 stars), Improve Codebase Architecture (ywwynm/EverythingDone, 144 stars) and Domain Modeling (brim-borium/spotify_sdk, 166 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
cosmicstack-labs (a GitHub organization) maintains it in cosmicstack-labs/mercury-agent-skills, which has 476 GitHub stars. The repository holds 12 skills in this directory. The repository was last updated on August 25, 2026.
Source: cosmicstack-labs/mercury-agent-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.