calibrate

calibrate is a command for Claude Code from let-sunny/canicode. It costs 0 tokens per session (550 once invoked), scanned A, original, MIT.

A calibration pipeline for testing how accurately a tool converts website fixtures into HTML. A fixture is a saved example used for repeatable testing.

In plain words
What is it for?
It can calibrate one fixture, all active fixtures, or resume a failed run while comparing outputs, evaluating gaps, updating rules, and producing an aggregate report.
Why use it?
It brings analysis, conversion, visual comparison, code measurements, review, and reporting into one repeatable run.

Command for Claude Code

Part of the canicode plugin — 3 skills, 3 commands, 3 hooks, 2 MCP servers shipped together

Install

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.

agentmods
npx agentmods add commands/let-sunny/canicode/calibrate
Clone the repo
git clone --depth 1 https://github.com/let-sunny/canicode

Made for: Claude Code.

Or install canicode, the plugin that ships this one along with the rest of its 3 skills, 3 commands, 3 hooks, 2 MCP servers.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for calibrate

README.md
[![agentmods](https://agentmods.dev/badge/commands/let-sunny/canicode/calibrate.svg)](https://agentmods.dev/commands/let-sunny/canicode/calibrate)
Your own site
<a href="https://agentmods.dev/commands/let-sunny/canicode/calibrate"><img src="https://agentmods.dev/badge/commands/let-sunny/canicode/calibrate.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 550 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00550
Opus 5 $0.00000 $0.00275
Sonnet 5 $0.00000 $0.00110
Haiku 4.5 $0.00000 $0.00055

Measured 5d ago against content hash 8fe46cc35a7c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

calibrate 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 5d 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.

.claude/commands/calibrate.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run the calibration pipeline for a single fixture, all active fixtures, or resume a failed run.

Input: $ARGUMENTS (fixture path, --all, or --resume <run-dir>)

Instructions

Run the calibration script:

npx tsx scripts/calibrate.ts $ARGUMENTS

The script (scripts/calibrate.ts) orchestrates:

  1. Analyze (CLI) — calibrate-analyze
  2. Design tree (CLI) — generate baseline + 6 stripped design-trees
  3. Convert (Agent, 7 parallel claude -p sessions) — baseline + 6 strip HTMLs
  4. Measure (CLI) — visual-compare + code-metrics for all 7
  5. Gap Analyze (Agent, claude -p) — if tier=full
  6. Evaluate (CLI) — calibrate-evaluate
  7. Critic (Agent, claude -p) — challenge proposals
  8. Arbitrator (Agent, claude -p) — final decisions, applies to rule-config.ts
  9. Evidence (CLI) — enrich + prune calibration evidence
  10. Report (CLI) — aggregate gap report

index.json tracks 14 internal sub-steps (e.g., screenshot capture, strip generation, post-processing split out separately). The 10 stages above are the logical grouping.

Modes

  • Single fixture: npx tsx scripts/calibrate.ts <fixture-path>
  • All active fixtures: npx tsx scripts/calibrate.ts --all — discovers active fixtures, runs sequentially, checks convergence, runs regression check, generates aggregate report
  • Resume: npx tsx scripts/calibrate.ts --resume <run-dir>

After the script completes

Report the summary:

  • Which steps completed, skipped, or failed
  • Similarity score (from the measure step summary)
  • Proposals and decisions (from evaluate/arbitrator summaries)
  • For --all mode: fixtures ran/passed/failed/converged
  • Path to the run directory (or aggregate report for --all)

Rules

  • The script handles all orchestration — do NOT manually run individual steps.
  • If the script fails, check $RUN_DIR/index.json for the failed step and error message.
  • To re-run from a failed step, use --resume. Do NOT delete the run directory.
  • Fixtures run sequentially in --all mode — each may modify rule-config.ts.

Read the full file on GitHub · 48 lines

Changes

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.

  1. 5d ago First seen · 48 lines · 0 tokens per session scan A 8fe46cc35a7c

Subscribe to this mod's changes

calibrate is a command published in the GitHub repository let-sunny/canicode (10 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 550 tokens. 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.