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 skills add XRenSiu/claude-code-forge --skill calibrategit clone --depth 1 https://github.com/XRenSiu/claude-code-forgeWrote 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.
[](https://agentmods.dev/skills/xrensiu/claude-code-forge/calibrate)<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/calibrate"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/calibrate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/calibrate"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/calibrate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00359 | $0.04921 |
| Opus 5.5 | $0.00144 | $0.01968 |
| Sonnet 5.5 | $0.00072 | $0.00984 |
| Haiku 4.5 | $0.00036 | $0.00492 |
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 1mo 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 — 255 lines — stays where its author put it; the contents beside it link to each section on GitHub.
calibrate
Prove that a compiled standard — an eval_case set or a rubric_version — is load-bearing,
not decorative, before it is allowed to gate anything. This is the meta-gate: the only
gate that judges the thing that judges the output. This skill describes the two mirrors that
certify a ruler, the two non-negotiables (holdout + isolation) that make the certification real,
and the meta-gate verdict that decides whether a standard may go live. It prescribes no step
order — the engine sequences the work; what follows are the gaps to fill and the gates that must
hold, in any order.
术语映射（在 AI-DLC 里怎么读这份文件）
本 skill 引自 qanat 仓库，正文保留其领域词汇；在 AI-DLC 里按下表读：
| 原文 | AI-DLC 里的对应物 |
|---|---|
| Territory（领地） | 一个 bounded context / 模块：dos.yaml 的 bounded_contexts.current_context；issue 的 Depends on DOS 所属上下文 |
| Run（一次执行） | 一次 issue → PR 的交付，即 .aidlc/<slug>/ 一个 slug |
| Contract / Contract 模板 | done_when.yaml（v2，以 AC 为单位）；模板 = 同类需求复用的 AC 骨架 |
| R001（评估者与执行者隔离） | AI-DLC 的信息隔离：实现子 agent 看不到评审判据；验收在独立会话 |
| R002（闸门资产只能人签） | AI-DLC 的 G2（判据冻结 lock_done_when.py sign --by <人>）与 G3（例外复核） |
| 变更提案 / G2 签字（AI-DLC） / change proposal / G2 signing (AI-DLC) / NEEDS_HUMAN | assets/change_proposal.md 变更提案 + G2/G3 人签；账本 ledger.md 记 propose |
verify_g1 / review_g2（qanat 的机器闸 / 评审闸） |
AI-DLC L7 的 A 档机械验收 / C 档判断验收——注意与 AI-DLC 的 G1（世界裁决）、G2（判据冻结）不是同一对门 |
| MemoryAsset（eval_case / rubric_version / failure_memory） | 归档目录 specs/<slug>/ 里的测试集 / 评判 rubric；failure_memory = ledger.md 的 fail 行 + escape-defects.md |
| daemon / 运行时 | 本地测试与 CI；/ai-dlc 的 acceptance 阶段 |
calibration.resolved 事件 |
G3 记录里"标准不清"的改判（g3_record.md） |
The gap (why a green test report is not a calibrated ruler)
A composite of three atoms: Judgment (what counts as a calibrated standard), Capability (run mutation testing / compute an agreement matrix / carve a holdout), and Control (the meta-gate seam — non-compliant forbids activation).
The load-bearing reason this is not free:
What ships with it
13 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- assets/calibration_report.yaml 6.2 KB
- assets/decisions_template.md 1.4 KB
- eval/fixtures/report_bad_noholdout.yaml 875 B
- eval/fixtures/report_good.yaml 1014 B
- eval/fixtures/report_holdout_dispute.yaml 1.4 KB
- eval/fixtures/report_holdout_impl_gap.yaml 1.2 KB
- eval/gate.json 3.8 KB
- eval/report.md 2.6 KB
- references/agreement-mirror.md 2.8 KB
- references/anti_patterns.md 3.4 KB
- references/holdout-and-isolation.md 3.5 KB
- references/mutation-mirror.md 2.9 KB
- scripts/verify_calibration.py 11 KB runs code
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.
- 1mo ago Changed · +18 lines d369fbf26914
- 1mo ago Changed b55cb448cea1
- 1mo ago First seen · 237 lines · 359 tokens per session scan A 61de0de30cba
calibrate is a skill published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed 1mo ago), licensed MIT. It adds 359 tokens to every session and 4,921 once invoked, about $0.0014 per session on Opus 5.5. 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 2026-09-05.
Other skills, from other repositories
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
tika-eval-compare
Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.
atmos-validation
Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.
skill-benchmark
Benchmark AI skill effectiveness by measuring implementation quality against legacy constraints.