agent-device-lab
Prepare and check trusted macOS hosts for physical Android and iPhone tests with agent-device, ADB, and Xcode, or set up iPhone…
Recover Codex or Claude sessions from tmux panes, saved cockpit snapshots, session history, and recent agent logs without overwriting useful restore evidence. Use when a pane crashes, detaches, loses visible context, needs a resume command, or requires a safe targeted restore
$ npx -y skills add vincentkoc/dotskills --skill codex-session-recovery --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/codex-session-recoveryContext preview
The summary Claude sees to decide when to auto-load this skill.
Recover Codex or Claude sessions from tmux panes, saved cockpit snapshots, session history, and recent agent logs without overwriting useful restore evidence. Use when a pane crashes, detaches, loses visible context, needs a resume command, or requires a safe targeted restore
name: codex-session-recovery description: Recover Codex or Claude sessions from tmux panes, saved cockpit snapshots, session history, and recent agent logs without overwriting useful restore evidence. Use when a pane crashes, detaches, loses visible context, needs a resume command, or requires a safe targeted restore plan. license: MIT metadata: source: "https://github.com/vincentkoc/dotskills"
Recover crashed, detached, or confusing agent sessions while preserving the best restore source.
1. Preserve evidence before writing state.
2. Inspect live tmux state.
3. Inspect available recovery sources.
4. Produce a dry-run plan first.
5. Execute only with clear intent.
6. Stay narrow for session-ID requests.
stateDiagram-v2
[*] --> PreserveRestoreEvidence
PreserveRestoreEvidence --> MatchSourceAndTarget
MatchSourceAndTarget --> ReportUnknown: identity or source ambiguous
MatchSourceAndTarget --> ReturnResumeID: session-ID lookup
MatchSourceAndTarget --> ShowExactPlan: preview or recovery
ShowExactPlan --> ReportPlan: preview only or mutation not authorized
ShowExactPlan --> TargetedRecovery: exact mutation authorized
TargetedRecovery --> VerifyTarget
VerifyTarget --> ReportResult
ReportUnknown --> [*]
ReturnResumeID --> [*]
ReportPlan --> [*]
ReportResult --> [*]We are moving from "prompt as text" to skill as runtime module. Each skill acts like a lightweight, containerized application for AI work: a stable interface, opinionated workflow, and bundled resources that can be installed, versioned, tested, and reused
Prepare and check trusted macOS hosts for physical Android and iPhone tests with agent-device, ADB, and Xcode, or set up iPhone…
Resolve canonical Git checkouts, index and verify codebase-memory-mcp graphs through the guarded CLI, open its browser viewer on…
Mine structured Codex /goal history locally or across a configured machine fleet, measure active goal time and resumed thread…
Interpret Crabpot and OpenClaw performance dashboard metrics, import-loop profiles, runtime profiles, and branch-to-branch perf…
Maintain and release the crawlkit Go library, preserving downstream compatibility for gitcrawl, slacrawl, discrawl, and notcrawl.
Template skill for repository authors; excluded from public publishing.