drift
AGENTS.md14 starsChanged 7 months agoMITView on GitHub
Agents, Skills und Evaluation-Prompts im Drift-Workspace
In this repository41 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .github/copilot-instructions.mdCopilot instructions
- llms.txtllms.txt
- brainstormingSkill
- drift-adr-workflowSkill
- drift-agent-prompt-authoringSkill
- drift-commit-pushSkill
- drift-dependency-updateSkill
- drift-docs-gap-fixSkill
- drift-evidence-artifact-authoringSkill
- drift-finding-message-authoringSkill
- drift-ground-truth-fixture-developmentSkill
- drift-mcp-repo-ops-usageSkill
- drift-pr-review-loopSkill
- drift-pr-reviewSkill
- drift-releaseSkill
- drift-risk-audit-artifact-updatesSkill
- drift-security-triageSkill
- drift-session-handover-authoringSkill
- drift-signal-development-full-lifecycleSkill
- fix-failing-actionsSkill
- grill-meSkill
- guard-src-drift-apiSkill
- guard-src-drift-commandsSkill
- guard-src-drift-ingestionSkill
- guard-src-drift-outputSkill
- guard-src-drift-signalsSkill
- guard-src-driftSkill
- improve-codebase-architectureSkill
- prd-to-issuesSkill
- self-improving-agentSkill
- simplify-and-hardenSkill
- skill-creatorSkill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- tddSkill
- test-driven-developmentSkill
- verification-before-completionSkill
- workspace-automation-orchestratorSkill
- write-a-prdSkill
- writing-plansSkill
The file
Files
Drift Agents & Prompts
Release Agent
Quick Release — Nach Code-Änderungen an src/drift/ den vollständigen Release-Workflow ausführen.
Verwendung
Im Chat erwähnen: „Release version", „Create release", „Publish to PyPI".
Der Agent: Validiert → Version berechnet → CHANGELOG + pyproject.toml → Commit + Tag → Push → PyPI.
Hinweis: Releases werden automatisch via python-semantic-release in CI verwaltet. Lokaler Fallback nur bei CI-Ausfall:
python scripts/release_automation.py --full-release
Wann verwenden
- Nach
feat:,fix:oderBREAKING:-Commits aufsrc/drift/ - Wenn Tests bestehen und Code committed ist
- Nicht bei unfertiger Arbeit oder fehlschlagenden Tests
Evaluation-Prompts
Prompts unter .github/prompts/ evaluieren die Drift-CLI aus verschiedenen Perspektiven.
Navigations-Guide: .github/prompts/README.md
| Prompt | Zweck | Empfohlene Reihenfolge |
drift-onboarding | First-Use-Erfahrung (Zero-Knowledge) | 1 |
drift-agent-workflow-test | Vollständiger CLI-Pfadtest | 2 |
drift-signal-quality | Signal-Precision/Recall messen | 3 |
drift-ci-gate | CI-Integration validieren | 4 |
drift-ai-integration | LLM-Context-Qualität bewerten | 5 |
drift-agent-ux | Agent-UX-Audit (Auffindbarkeit, Latenz, Fehler) | 6 |
PR-Orchestrator | PR-Bewertung aus Agent-Perspektive | nach Bedarf |
release | Release-Workflow validieren | nach Bedarf |
Shared Components
| Datei | Zweck |
_partials/bewertungs-taxonomie.md | Einheitliche Bewertungs-Labels |
_partials/konventionen.md | Gemeinsame Konventionen (Policy Gate, Pfade, Sandbox) |
_partials/issue-filing.md | Issue-Filing-Template (interne Prompts) |
_partials/issue-filing-external.md | Issue-Filing-Template (Field-Tests, Cross-Repo) |
Field-Test-Prompts (externe Repos)
Prompts unter .github/prompts/field-tests/ testen drift in beliebigen Repositories.
Issues gehen an mick-gsk/drift. Details: .github/prompts/field-tests/README.md
| Prompt | Zweck | Empfohlene Reihenfolge |
drift-field-test | Smoke-Test: Funktioniert drift? | 1 |
drift-finding-audit | Tiefenprüfung: Findings korrekt? | 2 |
drift-context-eval | Kontext-Qualität für AI-Workflows | 3 |
Skills
drift-agent-prompt-authoring— Drift-Prompts unter.github/prompts/undfield-tests/strukturiert schreiben oder verbesserndrift-release— Vollständiger Release-Workflowdrift-pr-review— PR-Review gemäß Drift Policydrift-security-triage— Security-Report-Triage
Dokumentation
- Release Skill:
.github/skills/drift-release/SKILL.md - Release Instructions:
.github/instructions/drift-release-automation.instructions.md - Policy:
.github/instructions/drift-policy.instructions.md - Push Gates:
.github/instructions/drift-push-gates.instructions.md
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.