create-egregore 0.19.5 → 0.19.7

package/runtime/pi/.codex/skills/invite/SKILL.md

@@ -0,0 +1,40 @@
+---
+name: invite
+description: Invite a GitHub user into this Egregore instance from Codex when the user invokes /invite or $invite, or asks to invite someone without exposing credentials.
+---
+
+# Egregore Invite
+
+Native Codex Egregore skill. Invite a GitHub username to the configured repos
+and, in connected mode, create the Egregore join link.
+
+## Flow
+
+1. Require a GitHub username. If missing, ask for it and stop.
+2. Read mode and config from `egregore.json`. Never print secrets from `.env`.
+3. Run credential-sensitive work in a single shell command that reads
+   `GITHUB_TOKEN` and `EGREGORE_API_KEY` internally and only prints sanitized
+   JSON.
+4. Connected mode:
+   - POST to the configured Egregore invite API with the GitHub username,
+     org, repo, slug, API key, and GitHub token.
+   - Parse the response and show invite status, memory access status, and the
+     returned invite URL.
+   - Best-effort graph record with `bin/graph.sh`.
+   - Best-effort direct notification with `bin/notify.sh send` only if a
+     contact channel exists.
+5. Local mode:
+   - Use GitHub API collaborator endpoints for the core repo, memory repo, and
+     managed repos.
+   - Create or update `memory/people/{username}.md`.
+   - Commit and push memory.
+   - Show the join command using `npx -y create-egregore@latest join`.
+6. If remote hosting is enabled, provision hosting best-effort in a detached
+   command and include the status only when it succeeds.
+
+## Rules
+
+- Never expose tokens, request bodies containing tokens, or raw API JSON.
+- Only org admins can complete the connected invite flow; report API denial
+  clearly and give the manual GitHub access URL.
+- Do not use Claude Code commands.

package/runtime/pi/.codex/skills/issue/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: issue
+description: 'Fallback Codex adapter for the Egregore issue workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore issue Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/issue/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/issue/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/me/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: me
+description: 'Fallback Codex adapter for the Egregore me workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore me Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/me/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/meeting/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: meeting
+description: 'Fallback Codex adapter for the Egregore meeting workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore meeting Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/meeting/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/meeting/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/minion/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: minion
+description: '>'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore minion Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/minion/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/note/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: note
+description: 'Fallback Codex adapter for the Egregore note workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore note Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/note/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/onboarding/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: onboarding
+description: 'Fallback Codex adapter for the Egregore onboarding workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore onboarding Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/onboarding/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/onboarding/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/pr/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: pr
+description: 'Fallback Codex adapter for the Egregore pr workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore pr Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/pr/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/project/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: project
+description: 'Fallback Codex adapter for the Egregore project workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore project Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/project/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/project/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/pull/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: pull
+description: 'Fallback Codex adapter for the Egregore pull workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore pull Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/pull/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/push/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: push
+description: 'Fallback Codex adapter for the Egregore push workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore push Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/push/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/quest/SKILL.md

@@ -0,0 +1,77 @@
+---
+name: quest
+description: Manage Egregore quests from Codex when the user invokes /quest or $quest, or asks to manage quests, using memory-first files and optional connected-mode graph updates.
+---
+
+# Egregore Quest
+
+Native Codex Egregore skill. Quests are shared, open-ended explorations stored
+under `memory/quests/`.
+
+## Flow
+
+1. Run `bin/agent.sh sync`.
+2. Detect mode from `egregore.json` `mode`, defaulting to `connected`.
+3. Dispatch from the user request:
+   - no argument: list active and paused quests from `memory/quests/*.md`.
+   - name or slug: show quest details from the markdown file.
+   - `new`: create a quest.
+   - `contribute`: append a contribution entry.
+   - `prioritize`: set `priority` to `0`, `1`, `2`, or `3`.
+   - `pause`: set `status: paused`.
+   - `complete`: set `status: completed` and record an outcome.
+4. Filesystem writes are canonical. Quest files use frontmatter:
+
+```yaml
+---
+title: Example Quest
+slug: example-quest
+status: active
+projects: []
+started: YYYY-MM-DD
+started_by: handle
+priority: 0
+completed:
+---
+```
+
+5. Update `memory/quests/index.md` after create, status changes, and priority
+   changes.
+6. Commit and push memory immediately:
+
+```bash
+git -C memory add quests
+git -C memory commit -m "quest: $SLUG $ACTION" --quiet
+git -C memory push origin main --quiet
+```
+
+7. In connected mode only, mirror changes best-effort with `bin/graph.sh` and
+   notify named people with `bin/notify.sh send`. If graph or notification
+   fails, report that the quest was saved to memory and continue.
+
+## Interaction
+
+Use structured Codex question tooling when available for new quest fields and
+status confirmations. Otherwise ask one short numbered question at a time with
+an `Other:` option.
+
+## Output
+
+Structured UX parity is required:
+
+- `/quest` list mode renders the active quest table:
+  `| Quest | Project | Artifacts | Contributors |`, followed by paused quest
+  state and entry-point hints.
+- `/quest <name>` detail mode renders the same sectioned view as Egregore:
+  title rule, status, projects, started metadata, question, threads,
+  artifacts, todos when available, contributors, and entry points.
+- Create/update routes use compact structured confirmations with the quest
+  file path and memory push state. Do not substitute a loose prose summary.
+- In local mode, omit graph/notification wording while preserving the same
+  list/detail layout.
+
+## Rules
+
+- Do not use graph or notification calls in local mode.
+- Do not let graph failure block a filesystem quest update.
+- Do not use Claude Code commands.

package/runtime/pi/.codex/skills/quest-suggest/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: quest-suggest
+description: 'Fallback Codex adapter for the Egregore quest-suggest workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore quest-suggest Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+1. Read `.claude/skills/quest-suggest/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/reflect/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: reflect
+description: 'Fallback Codex adapter for the Egregore reflect workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore reflect Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/reflect/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/reflect/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/review-pr/SKILL.md

@@ -0,0 +1,41 @@
+---
+name: review-pr
+description: 'Fallback Codex adapter for the Egregore review-pr workflow.'
+---
+
+<!-- generated-by: bin/codex-sync-skills.sh -->
+
+# Egregore review-pr Adapter
+
+This is fallback coverage for a long-tail Egregore workflow that has not been
+ported to a hand-written Codex-native skill yet.
+
+Use the project shell and filesystem directly. Do not invoke Claude Code
+commands. Translate interactive choices to structured Codex question tooling
+when it is available; otherwise render compact numbered choices with an
+`Other:` option and wait for the user.
+
+## Structured UX parity
+
+This workflow has a Claude skill with user-visible structured output. After
+reading `.claude/skills/review-pr/SKILL.md`, reproduce the same visible UX in Codex:
+
+- Preserve TUI boxes, markdown tables, rich cards, browser artifact rendering,
+  exact confirmation blocks, and "no preamble" rules from the source skill.
+- Use the source skill's frame width, section order, labels, status footer,
+  and examples as the contract for the final response.
+- Never replace a required box/table/card/artifact view with a prose summary
+  unless the user explicitly asks for a summary.
+- When the source says to output a TUI box directly, paste that box as the
+  visible response, preferably in a `text` fenced block.
+- Never show raw JSON, raw command output, or unformatted script output when
+  the source skill requires formatted status or rendered output.
+
+1. Read `.claude/skills/review-pr/SKILL.md` for the workflow details.
+2. Run the referenced `bin/` scripts directly from Codex.
+3. Treat graph, publish, and notify steps as best-effort unless that workflow
+   explicitly says they are required.
+4. Keep local-mode behavior filesystem-first and avoid graph or notification
+   calls when `egregore.json` declares `"mode": "local"`.
+5. Never call the deprecated `egregore-handoff` CLI for Egregore project
+   handoffs.

package/runtime/pi/.codex/skills/save/SKILL.md

@@ -0,0 +1,62 @@
+---
+name: save
+description: Save Egregore work when the user invokes /save or $save, or asks to commit, push, sync, or save current changes from a Codex Egregore session.
+---
+
+# Egregore Save
+
+Native Codex Egregore skill. This is the user-facing abstraction for the git
+and memory workflow: users should be able to say "save this" without knowing
+which branch, commit, push, pull request, or memory sync steps are required.
+
+## Flow
+
+1. Confirm this is an Egregore checkout by checking for `bin/agent.sh`.
+2. Inspect state silently:
+
+```bash
+git status --short
+git status --short memory 2>/dev/null || true
+git branch --show-current
+```
+
+3. If there are no repo or memory changes, say everything is already saved and
+   stop.
+4. Synthesize:
+   - a short topic from the work,
+   - a clear commit message,
+   - a one-line user-facing scope summary.
+5. If the scope is ambiguous or includes unrelated changes, ask for one compact
+   confirmation. Use structured Codex question tooling when available;
+   otherwise render:
+
+```text
+Save these changes?
+1. Save all
+2. Narrow scope
+Other:
+```
+
+6. Run the bridge command:
+
+```bash
+bin/agent.sh save --message "$MESSAGE" --topic "$TOPIC"
+```
+
+The bridge owns the mechanical workflow: sync memory, ensure a task branch,
+commit repo changes, push, and create or reuse a pull request when available.
+
+7. Parse the output and report only useful user information:
+   - branch name,
+   - whether a commit was created,
+   - whether push succeeded,
+   - pull request URL if present,
+   - memory sync status.
+
+## Rules
+
+- Do not ask the user to run git commands.
+- Do not expose implementation detail unless save fails.
+- If save fails, say which step failed and leave the branch/path clear.
+- Never run destructive git commands.
+- Do not use Claude Code commands.