Install
$ agentstack add skill-omexit-claude-skills-pack-adr ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
ADR (Architecture Decision Record)
What I'll do
Capture an architectural decision so future readers understand why the choice was made, what alternatives existed, and what trade-offs were accepted.
When to write an ADR
- Choosing between frameworks, libraries, or infrastructure
- Changing a data model, API contract, or system boundary
- Adopting or abandoning a pattern (monolith → microservices, REST → GraphQL)
- Any decision that would be hard to reverse and that a new team member would ask "why?"
How I'll think about this
- Context over conclusion: Spend more time explaining the forces at play (constraints, requirements, team capabilities, timeline) than the decision itself.
- Evaluate alternatives honestly: Each alternative should have real pros — not straw-man options designed to make the chosen one look good.
- Name the trade-offs: Every decision has costs. State them explicitly so future readers can re-evaluate if circumstances change.
- Keep it short: An ADR is a decision record, not a design doc. If it exceeds 2 pages, consider splitting out the detailed analysis.
Anti-patterns to flag
- Writing an ADR after implementation just to check a box (context is already lost)
- Only one alternative listed (signals the decision wasn't actually evaluated)
- Vague consequences ("this might cause issues later")
- Missing links to related ADRs or design docs
Quality bar
- A new team member can read this in 5 minutes and understand why the decision was made
- Alternatives have genuine pros, not just cons
- Consequences are concrete and specific
- Status field is set correctly (Proposed / Accepted / Superseded)
- Links to related ADRs or design docs if they exist
Workflow context
- Typically follows:
/design-doc(for decisions surfaced during design) - Feeds into:
/design-doc(as referenced context),/ticket-breakdown - Related:
/opportunity-assessment(strategic decisions)
Output
Create an ADR using templates/adr.md. Keep it short and specific.
Learning & Memory
After ADR creation completes, save:
- Decision patterns that recur across the codebase (technology choices, architectural styles, integration approaches)
- Alternatives considered and why they were rejected, for reuse in future similar decisions
- Context that informed the decision (constraints, team capabilities, timelines) to detect when circumstances change enough to revisit
Output contract
produces:
- type: "adr"
format: "markdown"
path: "claudedocs/-adr.md"
sections: [context, decision, alternatives, consequences]
handoff: "Write claudedocs/handoff-adr-.yaml — suggest: design-doc"
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: OmexIT
- Source: OmexIT/claude-skills-pack
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.