Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/adilkalam/orca/impeccablegit clone --depth 1 https://github.com/adilkalam/orcaWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00112 | $0.03672 |
| Opus 5 | $0.00056 | $0.01836 |
| Sonnet 5 | $0.00022 | $0.00734 |
| Haiku 4.5 | $0.00011 | $0.00367 |
Grade A, and why
impeccable scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/impeccable — Design Lane (three tiers)
The canon is ~/.claude/docs/reference/design-lane.md — the tier model, the precedence order
(owner > aesthetic > detector), the aesthetic-capture protocol, and the shared T1/T2 mechanics
(bind / judge / gate write / hook floor) are defined THERE, once. This command routes a request to a
tier and executes it. It restates nothing from the canon except the T0 banned-core (§2 —
deliberately inlined so T0 never depends on a skill load).
1. Routing
Explicit flags are deterministic:
| Argument | Tier |
|---|---|
--tweak |
T0 (§2) |
--layout --typeset --colorize --bolder --quieter --delight --harden --polish --optimize --adapt --clarify --distill --overdrive --animation --threejs |
T1 (§3) for that verb |
--full |
T2 (§4) |
--craft "<feature>" |
Feature setup (§6), then T2 |
--teach / --document / --extract |
Setup / maintenance (§6) |
Freeform (no flag): classify in one line, then go —
- a well-defined single-verb intent ("make this header breathe" -> quieter/typeset) -> T1;
- multi-verb / thin brief / whole-page ("clean up the pricing page") -> T2.
BIAS-TO-FAST (canon §Routing): when in doubt, route DOWN a tier. Wrong-fast is caught instantly by the owner's live eye on localhost; wrong-slow burns 20 minutes of machinery on a one-line change. The expensive failure is the slow one.
The standalone verb skills (/layout, /typeset, ...) are in-thread craft references WITHOUT the
judge or the detector floor — useful mid-conversation, but NOT an enforced path. Enforcement (fresh
judge + gate + hook floor) exists only here: /impeccable --<verb> (T1) or T2.
2. T0 — --tweak (in-thread, ZERO agent spawns)
Nudge a spacing value, swap one color token, fix one alignment seam. NO agent spawns, NO phase_state writes, NO gates, NO capture question. The owner is looking at the live surface; he is the verifier.
- The banned-core below is the floor — present because this command text carries it; do NOT rely on
Skill()loading. Loading~/.claude/skills/impeccable-hub/SKILL.mdis encouraged when depth is needed, but never assumed. - Make the edit(s) directly with
Edit/Write. - Run the detector inline as a REPORT on the changed files:
Report findings to the owner one line each, blocking findings named first. NEVER block, NEVER loop — the owner decides.node /Users/adilkalam/ORCA-OS/mcp/design-detector/bin/designcheck.js detect --json <changed files>
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 198 lines · 112 tokens per session scan A 8e6c8d86b796
impeccable is a command published in the GitHub repository adilkalam/orca (2 stars, last pushed 1mo ago), licensed MIT. It adds 112 tokens to every session and 3,672 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
profile
Save, load, inspect, update, reset, or delete diagram-design client profiles.
buildable-design
Create a UI/UX design brief from a prompt or the current Buildable app spec.
ui-flow-review
Review menus, HUD, navigation, and player flow from a UX perspective.
overlay
The spatial overlay — a click-through canvas floating over every window that both you and agents can draw on. /overlay shows status; /overlay open launches it (⌃⌥⌘D toggles draw mode, esc leaves); /overlay post --rel X,Y,W,H puts an arrow / box / highlight / spotlight / label / callout / sticker on the screen from an…
excalidraw-export
Export the current Excalidraw canvas scene, element set, library, or theme by retrieving the live resource snapshot.
design-engineer:discovery
Design workflow. For new products, runs the full pipeline. For existing projects, runs an abbreviated feature-focused flow. Argument feature-spec produces a truly minimal spec for established products with existing brand / design system.