calibrate-difficulty

A deterministic check that estimates which difficulty bands in an existing local AI run still have room for improvement. It uses a saved, hashed run artifact and avoids provider calls and holdout or customer data.

In plain words
What is it for?
Use it to classify sufficiently sampled bands as saturated, measurable, or needing investment, while keeping the data split and scoring source traceable.
Why use it?
It helps decide where training effort may be worthwhile without spending on new model runs or risking evaluation data.

Skill for Claude CodeCodex

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 skills/understudylabs/understudy-agent-tools/calibrate-difficulty
Any agent
npx skills add understudylabs/understudy-agent-tools --skill calibrate-difficulty
Clone the repo
git clone --depth 1 https://github.com/understudylabs/understudy-agent-tools

Made for: Claude Code, Codex.

Per session 24 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 374 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 $0.00024 $0.00374
Opus 5 $0.00012 $0.00187
Sonnet 5 $0.00005 $0.00075
Haiku 4.5 $0.00002 $0.00037

Measured 2d ago against content hash ad302ccaeb4d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

calibrate-difficulty 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.

skills/calibrate-difficulty/SKILL.md · 46 lines

What it actually says

Calibrate difficulty

Use an existing local run artifact with explicit generic band labels. The artifact is hashed and the report records that source binding. This is a deterministic screening step: it makes no provider calls, never reads holdout or customer data, and does not infer bands from benchmark-specific task IDs.

npm run build
node scripts/difficulty-calibration.mjs --run path/to/synthetic-run.json --out report.json

Scores at or above 0.95 are saturated only when the band has at least 10 scored rows; insufficient sample takes precedence and remains caution. Only sufficiently sampled bands below the threshold are measurable and invest. Keep fixture, split, scoring protocol, and source hash visible, and use dev/train data for decisions—never holdout data.

Safety Gates

  • Use only a frozen, source-bound train/dev run. Never use holdout rows or make provider calls from this calibration step.
  • Predeclare the saturation threshold and minimum sample. An undersized band is insufficient_sample even when its observed mean is perfect.
  • Treat the report as a spend-routing screen, not promotion evidence or proof that a model can beat the incumbent.

Resolve CLI

Build before invoking the repository-local script:

npm run build
node scripts/difficulty-calibration.mjs \
  --run path/to/source-bound-dev-run.json \
  --threshold 0.95 --min-sample 10 --out report.json

The script hashes the exact input bytes and writes a deterministic calibration report apart from generated_at.

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. 2d ago First seen · 46 lines · 24 tokens per session scan A ad302ccaeb4d

Subscribe to this mod's changes

calibrate-difficulty is a skill published in the GitHub repository understudylabs/understudy-agent-tools (16 stars, last pushed 3d ago), licensed MIT. It adds 24 tokens to every session and 374 once invoked, about $0.0001 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 2026-08-30.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

agent-host-chat-contributions

Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.

microsoft/vscode · 56 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens