drift: Skill for Claude Code

.github/skills/drift-evidence-artifact-authoring/SKILL.md

drift-evidence-artifact-authoring is a skill for Claude Code from mick-gsk/drift. It costs 108 tokens per session (3,210 once invoked), scanned A, original, MIT.

A workflow for creating versioned JSON evidence files that record benchmark results for feature changes in the Drift code-analysis project.

In plain words
What is it for?
Use it to produce benchmark evidence for new signals, heuristic changes, signal fixes, or reductions in false positives. It is intended for changes with passing tests and a known project version.
Why use it?
It removes uncertainty about the file name, structure, and contents needed to support a feature commit. It also checks that the evidence is generated in the required way.

Skill for Claude Code

Written for Claude Code: argument-hint in frontmatter.

This is mick-gsk/drift's own configuration. It tells Claude Code how to work on drift itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything drift configures →

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is python scripts/generate_feature_evidence.py --version X.Y.Z --slug my-feature.

Part of the drift plugin — 31 skills, 17 agents shipped together

Reuse

Borrowing it

Nothing to install: this file belongs to mick-gsk/drift. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/mick-gsk/drift/main/.github/skills/drift-evidence-artifact-authoring/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/mick-gsk/drift

Made for: Claude Code.

Or install drift, the plugin that ships this one along with the rest of its 31 skills, 17 agents.

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 drift-evidence-artifact-authoring

README.md
[![agentmods](https://agentmods.dev/badge/skills/mick-gsk/drift/drift-evidence-artifact-authoring/github.svg)](https://agentmods.dev/skills/mick-gsk/drift/drift-evidence-artifact-authoring)
Your own site
<a href="https://agentmods.dev/skills/mick-gsk/drift/drift-evidence-artifact-authoring"><img src="https://agentmods.dev/badge/skills/mick-gsk/drift/drift-evidence-artifact-authoring/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.

agentmods 80×15 button for drift-evidence-artifact-authoring

Your own site · 80×15
<a href="https://agentmods.dev/skills/mick-gsk/drift/drift-evidence-artifact-authoring"><img src="https://agentmods.dev/badge/skills/mick-gsk/drift/drift-evidence-artifact-authoring.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 108 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,210 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00108 $0.03210
Opus 5 $0.00054 $0.01605
Sonnet 5 $0.00022 $0.00642
Haiku 4.5 $0.00011 $0.00321

Measured 11d ago against content hash 40b8e0a97296, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

drift-evidence-artifact-authoring 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 11d 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.

.github/skills/drift-evidence-artifact-authoring/SKILL.md · 343 lines

How it starts

The opening of the file, as written. The whole thing — 343 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Drift Evidence Artifact Authoring Skill

Dieser Skill produziert ein vollständiges, commit-readies vX.Y.Z_<feature-slug>_feature_evidence.json in benchmark_results/.

Wann Verwenden

  • feat: commit geplant und drift-commit-push beanstandet fehlendes Evidence-File
  • Neues Signal oder materielle Heuristikänderung abgeschlossen
  • Precision-/Recall-Änderung > 5 Prozentpunkte vorhanden und zu belegen
  • Signal-Fix oder FP-Reduktion mit empirischen Belegen versehen

Pflicht-Vorbedingungen

  1. Tests laufen durch (make test-fast grün)
  2. ADR vorhanden oder Entscheidung explizit als fix/refactor eingestuft
  3. Version aus pyproject.toml oder geplanter nächster Version bekannt
  4. git diff --stat zeigt, welche Dateien tatsächlich geändert wurden

Step 1: Evidence-Datei deterministisch generieren (PRIMÄRE METHODE)

Ab Gate 2b ist manuelles Erstellen von Evidence-Dateien für neue feat:-Commits nicht zulässig. Der Pre-Push-Hook verifiziert, dass die Datei maschinell erzeugt wurde.

Befehl:

python scripts/generate_feature_evidence.py --version X.Y.Z --slug my-feature

Das Script:

  1. Läuft drift analyze --repo . --format json --exit-zero → erfasst drift_score, total_findings
  2. Läuft pytest tests/ -q --tb=short -m not slow -n auto → erfasst total_passing, total_failing
  3. Läuft pytest tests/test_precision_recall.py → erfasst Precision/Recall (optional via --skip-precision-recall)
  4. Bettet einen generated_by-Block mit git_sha, Timestamp und Script-Name ein
  5. Schreibt benchmark_results/vX.Y.Z_{slug}_feature_evidence.json

Flags:

--version X.Y.Z         Semver-Version (Pflicht)
--slug    my-feature    Kebab-Case-Slug (Pflicht)
--feature "Text"        Kurzbeschreibung (optional, Default: Slug)
--skip-tests            Nur Self-Analysis (kein pytest)
--skip-precision-recall Precision-Recall-Suite überspringen

Validierung manuell prüfen:

python scripts/validate_feature_evidence.py \
    benchmark_results/vX.Y.Z_my-feature_feature_evidence.json \
    --require-generated-by --push-head $(git rev-parse HEAD)

Read the full file on GitHub · 343 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. 11d ago First seen · 343 lines · 108 tokens per session scan A 40b8e0a97296

Subscribe to this mod's changes

drift-evidence-artifact-authoring is a skill published in the GitHub repository mick-gsk/drift (15 stars, last pushed 1mo ago), licensed MIT. It adds 108 tokens to every session and 3,210 once invoked, about $0.0005 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

contract-derivation

Use this skill after a design document exists and before any module is implemented. It takes a design doc and derives per-module contracts (manifests) that are complete enough for an implementor working in isolation to build each module correctly without seeing the rest of the system. Trigger this whenever the user…

taniwhaai/arai · 134 tokens

design-doc

Use this skill before writing any non-trivial code, especially when starting a new feature, service, or codebase. It produces a structural design document that commits to module boundaries, contracts, and data shapes BEFORE implementation begins. Trigger this whenever the user describes building something new, asks to…

taniwhaai/arai · 137 tokens

dispatcher

Use this skill when running a Taniwha build from the main Claude Code session. The skill turns the main session into a thin, mechanical executor that follows instructions from ephemeral orchestrator subagents. Trigger this whenever the user wants to start a Taniwha build, resume an interrupted one, or work on a…

taniwhaai/arai · 153 tokens

verifier

Use this skill when verifying that a completed Taniwha implementation or composition satisfies its contract's acceptance criteria. The verifier reads the contract, the implementation's source files, and the project context, then writes its own tests against the contract's acceptance criteria, runs them, and reports…

taniwhaai/arai · 122 tokens

composition

Use this skill when wiring two completed modules together to satisfy a parent contract, especially in a compartmentalised setting. Trigger this whenever the user has two implemented modules (each with its manifest) and a parent contract that says how they should compose, when the user asks to "compose these modules"…

taniwhaai/arai · 136 tokens

leaf-implementation

Use this skill when implementing a single module against a contract manifest, especially in a compartmentalised setting where the implementor should NOT see the rest of the system. Trigger this whenever a manifest exists and the user wants to implement that module, asks to "implement this module", "build against this…

taniwhaai/arai · 133 tokens