architect
Sketch caller usage, types, signatures, ownership, and module boundaries before implementation.
Skill by kyeshmz
Chalk Labs
Need help deploying architect?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
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/kyeshmz-architect
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: design
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
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/architect
What's inside
- Skill name
architect- Description
- Sketch caller usage, types, signatures, ownership, and module boundaries before implementation. Compare distinct shapes and revisit a sketch that produces repeated workarounds. Use for @architect, architect this, design this, or a nontrivial architecture decision.
- License
- MIT
- Metadata
- opencode/autoinvoke: false
- Instructions
- 52 lines in SKILL.md · view on GitHub
Files (6)
- LICENSE
- SKILL.md
- design-red-flags.md · references/
- rationale-template.md · references/
- runner-prompt.md · references/
- shared-rules.md · references/
SKILL.md (first lines)
# Architect Load `code-organization` as the canonical owner of design rules. This compatibility entry point adds candidate synthesis, a design rationale, an optional checkpoint, and the loop from implementation back to the sketch. It does not require an agent named `architect`. Read `references/shared-rules.md`. Resolve bundled files from this skill's loaded base directory. Work in the target repository. Keep a design-only request read-only apart from explicitly requested design artifacts. A request to sketch does not authorize implementation, commits, or external writes. Track Ground, Sketch, Agree, Implement, and Reconsider in a task list or short checklist. Skip implementation phases when the request ends at design. ## Ground 1. Load `how` for each affected subsystem. Establish runtime flow, callers, types, state owners, and integration constraints. A file list is not grounding. Reuse evidence already gathered for this task. 2. Load `why` when changing ownership or layering requires the existing rationale. Treat established constraints as design inputs, not guesses. 3. Skip existing-system grounding only for genuinely greenfield work with no surrounding integration. ## Sketch 1. Write README-style caller usage and two or three realistic call sites first. Derive types and signatures from that usage. 2. Compare at least two structurally distinct candidates. Whole shapes count. Small variations of one shape do not. Load `arena` for independent candidate generation and synthesis when useful. A local comparison is acceptable when delegation is unavailable, but it is not independent review. 3. Give each candidate the grounding and `references/runner-prompt.md`. Fill in the task, absolute working directory, owned output paths, constraints, and required evidence. Candidates produce sketches, not production changes, and do not launch their own workers. 4. Screen every candidate with `references/design-red-flags.md` and the canonical organization rules. Compare meaningful complexity hidden behind each public surface, not implementation ease alone. Prefer a change that is correct for the repository when a maintainer sees only one local example. 5. Select one coherent base and adapt worthwhile parts of the others. Record the choice, rejected shapes, grafts, judge availability, and verification in a task-local rationale following `references/rationale-template.md`. Never edit the bundled template to record a run. Inherit the current model unless the user authorizes an override and the host confirms its exact ID. No separate role file is required. Use one type-and-signature sketch for a small change. Use a module map plus core types for larger work. Mark sketch bodies `not implemented` and use brief pseudocode only where it clarifies a hard decision. Keep scaffolds isolated from runnable code unless the user requested that integration. ## Agree Proceed within the requested scope by default. Pause after synthesis when the user asks for a checkpoint or an unresolved product preference blocks the design. Present the usage, shape, tradeoffs, and open questions before asking for a decision. For authorized implementation work, the sketch is the contract. A separate scaffold commit is optional and requires commit authorization. Load `interrogate` if adversarial pressure on the sketch is requested or warranted. Human pushback is new grounding evidence. Re-ground and compare shapes again rather than defending a stale sketch. ## Implement Replace sketch bodies with code and verify each unit. Surface deviations instead of silently expanding the contract. For a missing parameter or new exception, decide whether the sketch missed a constraint, the requirement changed, or the implementation exceeds scope. ## Reconsider Revisit the architecture when implementation shows a pattern, not merely one hard edge case. - The same workaround appears in unrelated code. - Unrelated edge cases need the same special branch. - Types need `any`, casts, or optional fields that are always present in practice. - Shared writers require a lock where the sketch promised isolated state. - Callers must learn private rules to use the interface. - At least two independent implementation deviations have the same shape. Complex data alone does not condemn a design. When the pattern is real, load `how` over the implementation, then `principle-redesign-from-first-principles` and `principle-subtract-before-you-add`. Redesign with the new constraints treated as original requirements. Return to Sketch. Replace only artifacts this task owns. Preserve unrelated work and do not reset a checkout to discard a sketch. ## Report Return the usage-first sketch, module map when needed, and completed rationale. Name the chosen shape, alternatives, interface depth, synthesis decision, accepted tradeoffs, open questions, and first implementation step. Include checks actually run and any unavailable independent comparison. For implementation work, add deviations and behavior verification.
From kyeshmz/jiro, MIT.
More from kyeshmz/jiro
- 0.570score
create-formal-verification-skill
kyeshmz
Generate and prove a project-local formal verification skill for a high-risk invariant, race, parser, retry, or unsafe operation.
SkillAFreshness: AClaude Code +9 - 0.570score
jiro
kyeshmz
Jiro's verified engineering workflow for OpenCode.
SkillAFreshness: AClaude Code +9 - 0.567score
how
kyeshmz
Use for how does X work, code walkthroughs before changing something, and placement, ownership, or layering questions such as where should this live or which…
SkillAFreshness: AClaude Code +9 - 0.567score
interrogate
kyeshmz
Challenge code or a design with independent adversarial reviewers and synthesize a pragmatic verdict without applying fixes.
SkillAFreshness: AClaude Code +9 - 0.567score
arena
kyeshmz
Generate independent candidates for one task, score them against a private rubric, select a base, and synthesize the strongest coherent result.
SkillAFreshness: AClaude Code +9 - 0.567score
automate-me
kyeshmz
Turn the user's working preferences into a personal style skill or update one from explicit feedback and scoped conversation evidence.
SkillAFreshness: AClaude Code +9 - 0.567score
benchmark-checklist
kyeshmz
Vet a performance claim against its limiter, production tuning, physical limits, errors, repeatability, relevance, and proof that timed work happened.
SkillAFreshness: AClaude Code +9 - 0.567score
blast-radius
kyeshmz
Trace what a change can break beyond its diff and prove the load-bearing safety fact against real code.
SkillAFreshness: AClaude Code +9 - 0.567score
create-subagent
kyeshmz
Create or update a reusable OpenCode subagent with a clear role, required skills, permissions, and reporting contract.
SkillAFreshness: AClaude Code +9 - 0.567score
delegate
kyeshmz
Scope work into bounded child-agent tasks, write self-contained briefs, and verify their results.
SkillAFreshness: AClaude Code +9 - 0.567score
figure-it-out
kyeshmz
Design and execute an auditable workflow when no narrower recipe fits a large migration, multi-part change, or unattended task.
SkillAFreshness: AClaude Code +9 - 0.567score
ketchup
kyeshmz
Catch the user up on changes in this chat since their previous message and list every decision waiting on them with enough context to decide.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 19, 2026 (11 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/kyeshmz/jiro
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from kyeshmz
- 0.570score
jiro
kyeshmz
Jiro's verified engineering workflow for OpenCode.
SkillAFreshness: AClaude Code +9 - 0.570score
create-formal-verification-skill
kyeshmz
Generate and prove a project-local formal verification skill for a high-risk invariant, race, parser, retry, or unsafe operation.
SkillAFreshness: AClaude Code +9 - 0.567score
how
kyeshmz
Use for how does X work, code walkthroughs before changing something, and placement, ownership, or layering questions such as where should this live or which…
SkillAFreshness: AClaude Code +9 - 0.567score
arena
kyeshmz
Generate independent candidates for one task, score them against a private rubric, select a base, and synthesize the strongest coherent result.
SkillAFreshness: AClaude Code +9 - 0.567score
automate-me
kyeshmz
Turn the user's working preferences into a personal style skill or update one from explicit feedback and scoped conversation evidence.
SkillAFreshness: AClaude Code +9 - 0.567score
interrogate
kyeshmz
Challenge code or a design with independent adversarial reviewers and synthesize a pragmatic verdict without applying fixes.
SkillAFreshness: AClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get architect running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
architect skill. Agent Mart. https://agentmart.dev/skills/kyeshmz-architect (last updated 2026-10-24)