mortennordbye/agentic-dr/dry-run
Claude Code and agent skill by mortennordbye/agentic-dr
- community
Checking that the SKILL.md file is still in the repository.
Regenerate the Phase-2 disaster-recovery Terraform estate from current main and check it — generate + invariant-lint + reconcile + validate, plus the platform/core GitOps overlay, with an optional gated plan-only drift check (ARCHITECTURE.md Mode 1). Never applies. Enforces the hard rule: NEVER trigger a pipeline (even plan_only) without explicit user approval. Use when asked to dry-run, rehearse, test, drift-check or just generate the DR estate without touching cloud.
Install
npx vskill@latest install mortennordbye/agentic-dr/dry-runSKILL.md
/agentic-dr:dry-run — generate and check the DR estate (Mode 1)
Regenerates the DR estate from current main and proves it holds together, without ever applying anything. This is the mode to run constantly: it is cheap, it is the drift baseline, and it is what must be trusted before /agentic-dr:failover is ever reached.
Plugin root: ${CLAUDE_PLUGIN_ROOT} — this line is substituted, so that is a real absolute path. Use it wherever the procedure below writes <plugin-root>.
Procedure
- Run Phases 1–3 exactly as written in
${CLAUDE_PLUGIN_ROOT}/docs/build-procedure.md— read it
now, in full. It holds the golden rules, Phase-1 discovery and the build-plan approval gate, the Workflow fan-out, and the GitOps residue you own. Everything there applies here unchanged.
- Then stop, or escalate to a gated drift check. Generation + lint/reconcile/validate is a
complete Mode-1 run; you may end there. If the engineer wants a drift check against real cloud state: [GATE] ask before gh workflow run <DR pipeline> -f plan_only=true ..., then run it in dependency order, as far as a cold region allows (cold plan-only is partial by design, §6.3).
- No apply, ever.
plan_only=falseis not a Mode-1 outcome. If the run reveals something that
needs applying, that is a /agentic-dr:failover decision the engineer makes deliberately.
What is different from a failover
- GitOps sentinels stay unresolved. The overlay +
<state-dir>/gitops-report.mdare the artifact —
review the diff (git diff --no-index <source_dir> <target_dir>). __DR_POST_APPLY__* values do not exist until the producing roots apply, so do not deploy this overlay and do not hand-fill them. The failover skill carries the sentinel gate; here their presence is expected.
- The PR is the baseline, not a deployment. Committing the generated roots + state files is what
makes the next run's diff meaningful (ARCHITECTURE §16.2).
Report as described in docs/build-procedure.md — the pinned SHA, in-scope list, built vs blocked, open blackboard items and cycles, the manual-work checklist, and the PR link. Say plainly that nothing was applied.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/dry-run/SKILL.md
Popularity
- Repository stars
- 1
- 7-day trend score
- 3
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mortennordbye/agentic-dr/dry-run)