本文へ移動
cccskills
無料GitHub で公開

refocusing

Use when a long-running agent may have lost the product outcome, when context was compacted, when work has crossed a major boundary, or when a fresh path-based trace is needed before the next consequential action. Not for fresh-session orientation, waking an idle seat, or checkpoints.

インストール方法を見る

含まれるファイル(3)

  • SKILL.md3.4 KB
  • references/refocus.md1.1 KB
  • scripts/trace-to-root.py24.0 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Refocusing

Invoke the skill as refocusing, its frontmatter name. OpenRig seeds it in the Claude and Codex global skill roots and includes selected plugin skills in managed loadouts. Existing externally managed copies remain authoritative. If the harness has not discovered it yet, read this installed skill file and run its script directly; hook registration alone does not prove native skill discovery.

Refocus preserves a long session's earned expertise while re-grounding it in current intent and lived context. It is not a restart, wake, or phase checkpoint.

During an actionable managed restore, consume the current topology and work trace that actually arrived with the request. Do not rerun Python merely to duplicate it. If no current trace arrived, name that delivery gap. A packet pointer, compact summary or truncated extract is not a full source read. Use the native file-read tool for required notes and full sources, and complete the existing restore audit; partial file reads do not establish completed refocus or restoration.

For a new trace outside that delivered context, follow the command guidance below. Resolve the bundled script's absolute path from the directory of this loaded skill. Stay in your current working directory; do not cd into the skill directory, which can change the inferred work node. Run the trace as one plain command, replacing the example path with the actual absolute path written out in the tool call:

python3 "/absolute/path/to/refocusing/scripts/trace-to-root.py" --trees both --depth light

Do not combine this command with shell variables, environment assignments, command substitution, exit-status printing or file-reading loops. Read the named source files separately with the native read tool. Combined shell loops or variables can require native approval even when the intended operations are read-only; a plain command is not a guarantee that approval will never be needed.

For managed compaction, refocus during the restore request or its read-depth audit. The earlier acknowledgement-only boundary is not permission to restore.

Use --trees topology|work|both to select context domains and --depth light|full to control how much each node contributes. Light work traces compose intent: and name notes; full traces include the complete node and notes bodies. The script resolves topology.root and workspace.root with rig config get. When a known current node must be supplied, pass --work-start or --topology-start with its literal absolute path in the same plain command. Do not substitute the project root for a mission or slice node and claim that mission was traced.

The trace walks the topology and work trees. For the project's own declared context (intent, context files, skills), rig context work-install lists it; read what the next action needs.

Read references/refocus.md when changing the automatic hook or its content ladder. A missing chain file is evidence: report the gap and continue; never follow pointers to invent a second parent.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

Use when designing, building, operating, or diagnosing an ongoing application whose live backend or control loop includes OpenRig agents, including applications with a Markdown, YAML, or JSON agent control plane or a thin surface over specialist agent roles.

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

Use when a bounded real-world procedure has a deterministic happy path but brownfield, variable, or partially knowable state; when an operation must resume from verified evidence; or when deciding whether agent judgment or ordinary code should own a procedure's control loop.

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

Use when creating, refreshing, packaging, inspecting, promoting, or deprecating a named per-seat starting point — Agent Starter manifest authoring, the 6-state lifecycle (captured → named → inspectable → used → promoted → deprecated), provenance honesty, and refusal rules. NOT a VM image; a managed starting point composed from agent role + startup context + optional native session source + provenance.

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

Use when designing or auditing how an agent becomes useful after launch — AGENTS.md overlays, role files, skills, rig specs, workflow specs, startup checklists, refocus messages, "rig context" surface. Covers the 4 failure modes that make startup context fail (old rig spec misses current operating mode; current agents never told about new guidance; startup file as dumping ground; orchestrator transmits implementation without preserving product intent).

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

Use before launching a team during agent-guided setup, or when a user asks to configure OpenRig command permissions, reduce repeated native approval prompts, or apply a selected rig/seat permission policy.

日本語の概要は準備中です。原文の説明を表示しています。

mvschwarz/openrig6,8042026年10月11日 更新

mvschwarz のスキルをすべて見る

このスキルの問題を報告する