Compare commits

..

9 Commits

Author SHA1 Message Date
Jesse Vincent
2b205cf0e1 Merge pull request #1124 from obra/add-worktree-consent-step
enhance worktree consent during implementation (PRI-974)
2026-04-13 16:49:55 -07:00
Drew Ritter
884b1a5960 fix: treat direct worktree skill invocation as consent (PRI-974)
The skill previously required an explicit reply to its "do you want a
worktree?" dialogue, which produced obtuse UX when the user invoked the
skill by name — agents had to stop and ask "do you want a worktree?"
even though the user just asked for the skill whose purpose is worktrees.

Loosen Step 2 to recognize the invoking turn as consent: if the user's
most recent message named the skill, asked for a worktree, or asked for
an isolated workspace, proceed directly to Step 3 without re-prompting.
The gate still fires for the transitive case (agent infers isolation
from a feature description) — that remains the #991 failure mode.

Also trim "or skill invocation" from the anti-inference Red Flag and
destale the Integration section now that SDD/executing-plans no longer
require a worktree.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:31:10 -07:00
Drew Ritter
018f3675e5 fix: update stale references and restore silence safety net (PRI-974)
Post-inversion cleanup:

- executing-plans, subagent-driven-development: update Integration
  description from "Ensures isolated workspace" to "Detects workspace
  environment and offers worktree isolation on request"
- codex-tools.md: update step references (Step 0→1, Step 1→2)
- using-git-worktrees Step 2: restore "silence → ask once more" instead
  of "silence → work in place" to preserve safety net for confused users

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:31:10 -07:00
Drew Ritter
7c4597af34 fix: invert worktree skill default to work-in-place, eliminating Step 0.5 (PRI-974)
Agents consistently skipped Step 0.5 (consent gate) because fractional
numbering signals "optional afterthought" and the prose-only step was
invisible to code-block anchoring. The fix inverts the structural
gravity: the default path now works in place, and worktree creation
is an off-ramp requiring explicit user request.

- Renumber to clean integers: Step 1 (detect) → 2 (offer) → 3 (create) → 4 (setup) → 5 (verify)
- Step 2 defaults to Step 4 (in-place); Step 3 only on explicit user ask
- Step 2 includes a code block so agents register it during execution
- Add "creating without being asked" to Common Mistakes
- Add anti-inference red flag: consent from task/plan/skill invocation doesn't count

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:31:10 -07:00
Drew Ritter
b0e08a497f fix: promote consent to own Step 0.5 with structural enforcement (PRI-974)
Drill benchmark showed 0/4 consent compliance across both Claude Code
and Codex. Root cause: consent was buried inline in Step 0's conditional
branch. Agents anchor on the next bash command and skip prose.

Fix: promote consent to its own numbered section with imperative framing
("REQUIRED STOP", "Do NOT proceed without an answer") and exact output
template. Also adds explicit "no" path — users who want to work directly
on their current branch skip to Step 3 with no worktree creation.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:31:10 -07:00
Drew Ritter
ddbba8e469 docs: drop brittle step-number chain from multi-repo row
Addresses review feedback on #1123. Replaces "(same Step 0→1a→1b flow,
matching branch names)" with plain-language instruction that doesn't
forward-reference section numbers that could rot under future edits.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:30:00 -07:00
Drew Ritter
f0728841d8 feat: add multi-repo worktree guidance (#710) 2026-04-13 16:29:59 -07:00
Drew Ritter
e3dd3b4c5a fix: replace hardcoded /Users/jesse with generic placeholders (#858) 2026-04-13 16:29:59 -07:00
Drew Ritter
e4a15b6d52 docs: drop instruction file enumeration per PR #1121 review
Jesse flagged that the verbose CLAUDE.md/AGENTS.md/GEMINI.md/.cursorrules
enumeration (a) chews tokens, (b) confuses models that anchor on exact
strings, and (c) is repeated DRY-violatingly across 3+ locations.

Replace with abstract "your instructions" framing in four spots:
- skills/using-git-worktrees/SKILL.md Step 0 → Step 1 transition
- skills/using-git-worktrees/SKILL.md Step 1b Directory Selection
- docs/superpowers/plans/2026-04-06-worktree-rototill.md (both mirror locations)

Same intent, harness-agnostic phrasing, ~half the tokens.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-13 16:29:44 -07:00
2 changed files with 10 additions and 8 deletions

View File

@@ -251,7 +251,7 @@ Report with branch state:
**If `GIT_DIR == GIT_COMMON` (or in a submodule):** You are in a normal repo checkout. **If `GIT_DIR == GIT_COMMON` (or in a submodule):** You are in a normal repo checkout.
If the user has not already made their preferences about worktree isolation clear — check global and project agent instruction files (`CLAUDE.md`, `AGENTS.md`, `GEMINI.md`, `.cursorrules`, or equivalent) — ask for consent before creating a worktree: Has the user already indicated their worktree preference in your instructions? If not, ask for consent before creating a worktree:
> "Would you like me to set up an isolated worktree? It protects your current branch from changes." > "Would you like me to set up an isolated worktree? It protects your current branch from changes."
@@ -289,7 +289,7 @@ Follow this priority order:
``` ```
If found, use it (backward compatibility with legacy global path). If found, use it (backward compatibility with legacy global path).
3. **Check your project's agent instruction file** (CLAUDE.md, GEMINI.md, AGENTS.md, .cursorrules, or equivalent) for a worktree directory preference. If specified, use it without asking. 3. **Check your instructions for a worktree directory preference.** If specified, use it without asking.
4. **Default to `.worktrees/`.** 4. **Default to `.worktrees/`.**

View File

@@ -48,7 +48,9 @@ echo "Current branch: $BRANCH"
echo "Repository: $(basename "$(git rev-parse --show-toplevel)")" echo "Repository: $(basename "$(git rev-parse --show-toplevel)")"
``` ```
Tell the user their options, then **wait for a reply**: **Check the user's most recent message first.** If they already asked for a worktree, named the worktree skill, or asked for an isolated workspace in the message that invoked you, that IS the explicit ask — proceed directly to Step 3 without re-prompting.
Otherwise, tell the user their options and **wait for a reply**:
> "You're on `<branch>` in `<repo>`. I can set up an isolated worktree, or we can work directly here. What do you prefer?" > "You're on `<branch>` in `<repo>`. I can set up an isolated worktree, or we can work directly here. What do you prefer?"
@@ -82,7 +84,7 @@ Only proceed to Step 3b if you have no native worktree tool available.
Follow this priority order. Explicit user preference always beats observed filesystem state. Follow this priority order. Explicit user preference always beats observed filesystem state.
1. **Check global and project agent instruction files** (`~/.claude/CLAUDE.md`, `~/.config/gemini/AGENTS.md`, project-level `CLAUDE.md`, `AGENTS.md`, `GEMINI.md`, `.cursorrules`, or equivalent) for a declared worktree directory preference. If specified, use it without asking. 1. **Check your instructions for a declared worktree directory preference.** If the user has already specified one, use it without asking.
2. **Check for an existing project-local worktree directory:** 2. **Check for an existing project-local worktree directory:**
```bash ```bash
@@ -241,7 +243,7 @@ Ready to implement <feature-name>
- Create worktree without verifying it's ignored (project-local) - Create worktree without verifying it's ignored (project-local)
- Skip baseline test verification - Skip baseline test verification
- Proceed with failing tests without asking - Proceed with failing tests without asking
- Infer worktree consent from the task description, plan, or skill invocation — only an explicit user reply counts - Infer worktree consent from the task description or plan — only an explicit user request counts
**Always:** **Always:**
- Run Step 1 detection first - Run Step 1 detection first
@@ -256,9 +258,9 @@ Ready to implement <feature-name>
## Integration ## Integration
**Called by:** **Called by:**
- **subagent-driven-development** - Ensures isolated workspace (creates one or verifies existing) - **subagent-driven-development** - Calls this to detect the workspace and optionally set up worktree isolation on request
- **executing-plans** - Ensures isolated workspace (creates one or verifies existing) - **executing-plans** - Calls this to detect the workspace and optionally set up worktree isolation on request
- Any skill needing isolated workspace - Any skill that may use worktree isolation
**Pairs with:** **Pairs with:**
- **finishing-a-development-branch** - REQUIRED for cleanup after work complete - **finishing-a-development-branch** - REQUIRED for cleanup after work complete