fix: Deduplicate and restructure pause-point skill prose#1975
Merged
hatayama merged 2 commits intoJul 24, 2026
Conversation
The pause-point SKILL.md had grown to 29.3KB across ten feedback rounds, with the same facts restated in two to four places (pre-line snapshot semantics, --resume-play behavior, domain-reload clearing, JIT inlining, Step-after-hit) and field-log style narration in the physics-callback caveat. Since the skill body loads into agent context on every trigger, this size is a per-use token cost. Compress it to 20.2KB without dropping any contract fact: - State each fact once at a canonical location and remove restatements, historical rationale, and sales prose - Move failure-time-only content to a new references/troubleshooting.md (timeout diagnosis fields, entry-line bisection, JIT inlining, physics-callback misses, pre-bound delegates, sequence-point errors) - Move the self-progressing-game workflow details to a new references/fast-progressing-games.md (ResumePlayResult semantics, Time.timeScale warning, residual race, paused-state injection) - Move interpretation details (Step sessions, Rigidbody mid-solver values, preview budget scaling, raw capture tuple usage) into the existing references/captured-variables.md References load lazily, so the hot path shrinks by ~9KB while every fact stays reachable exactly once.
Contributor
|
Warning Review limit reached
Next review available in: 51 minutes Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (9)
📝 WalkthroughWalkthroughThis documentation-only change aligns pause-point command semantics, captured-variable interpretation, marker workflows, fast-progressing-game procedures, and troubleshooting guidance across the agent, Claude, and package skill references. ChangesPause-point documentation
Estimated code review effort: 2 (Simple) | ~10 minutes Sequence Diagram(s)sequenceDiagram
participant CLI
participant PausePoint
participant PlayMode
CLI->>PausePoint: arm marker
PausePoint-->>CLI: confirm arming
CLI->>PlayMode: resume when requested
CLI->>PausePoint: dispatch trigger
PausePoint-->>CLI: return hit and captured evidence
Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This was referenced Jul 24, 2026
Contributor
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.agents/skills/uloop-pause-point/references/fast-progressing-games.md:
- Line 35: Separate direct-write verification from post-Step behavior in
fast-progressing-games.md: immediately reread the written value while paused to
verify the injection, then optionally run control-play-mode --action Step and
document the resulting value separately because Update may recompute it. Apply
this clarification at
.agents/skills/uloop-pause-point/references/fast-progressing-games.md lines
35-35, .claude/skills/uloop-pause-point/references/fast-progressing-games.md
lines 35-35, and
Packages/src/Editor/CliOnlyTools~/PausePoint/Skill/references/fast-progressing-games.md
lines 35-35.
In
`@Packages/src/Editor/CliOnlyTools`~/PausePoint/Skill/references/captured-variables.md:
- Line 84: Update the captured-variables guidance at
Packages/src/Editor/CliOnlyTools~/PausePoint/Skill/references/captured-variables.md:84
to explicitly resume PlayMode or pass --resume-play before sending simulated
input, then regenerate the same instruction in
.agents/skills/uloop-pause-point/references/captured-variables.md:84 and
.claude/skills/uloop-pause-point/references/captured-variables.md:84.
In `@Packages/src/Editor/CliOnlyTools`~/PausePoint/Skill/SKILL.md:
- Line 132: Update the mirrored pause-point examples to use the fully qualified
UnityEngine.Debug.Assert assertion. Apply the change at
Packages/src/Editor/CliOnlyTools~/PausePoint/Skill/SKILL.md:132-132, then
regenerate or update the copies at
.agents/skills/uloop-pause-point/SKILL.md:132-132 and
.claude/skills/uloop-pause-point/SKILL.md:132-132.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: ee88dd4c-a1d6-4da8-9fa0-b9650b31b584
📒 Files selected for processing (12)
.agents/skills/uloop-pause-point/SKILL.md.agents/skills/uloop-pause-point/references/captured-variables.md.agents/skills/uloop-pause-point/references/fast-progressing-games.md.agents/skills/uloop-pause-point/references/troubleshooting.md.claude/skills/uloop-pause-point/SKILL.md.claude/skills/uloop-pause-point/references/captured-variables.md.claude/skills/uloop-pause-point/references/fast-progressing-games.md.claude/skills/uloop-pause-point/references/troubleshooting.mdPackages/src/Editor/CliOnlyTools~/PausePoint/Skill/SKILL.mdPackages/src/Editor/CliOnlyTools~/PausePoint/Skill/references/captured-variables.mdPackages/src/Editor/CliOnlyTools~/PausePoint/Skill/references/fast-progressing-games.mdPackages/src/Editor/CliOnlyTools~/PausePoint/Skill/references/troubleshooting.md
- Qualify the inline conditional-block example as UnityEngine.Debug.Assert so the example itself cannot be read as System.Diagnostics.Debug.Assert - State the resume step explicitly in the self-progressing-game workflow: simulate-* input requires an unpaused PlayMode, so arm the next marker with --resume-play or run control-play-mode --action Play before sending the verifying input - Split paused-state injection verification into two stages (immediate re-read to confirm the write landed, post-Step re-read to confirm it survives a frame) so a failed write is distinguishable from an Update() recompute
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The pause-point skill body had grown to 29.3KB (~7k tokens) across ten verification-feedback rounds, with the same facts restated in 2–4 places and field-report narration in the physics-callback caveat. The skill body loads into agent context on every trigger, so this is a per-use token cost.
This PR compresses it to 20.2KB (-31%) as a fact-preserving edit: every deleted sentence's fact either remains stated once at a canonical location in SKILL.md or moves to a lazily-loaded reference.
--resume-play, domain-reload clearing, JIT inlining, Step-after-hit, manual-pause countdown) and drop historical rationale / sales prosereferences/troubleshooting.md: timeout diagnosis fields, entry-line bisection, JIT inlining, physics-callback misses (rewritten from field-log style into symptom → confirmation → recovery order), pre-bound delegates, sequence-point enable failuresreferences/fast-progressing-games.md: freeze→build→resume workflow,ResumePlayResultsemantics,Time.timeScalewarning, residual race handling, paused-state injection caveatsreferences/captured-variables.mdabsorbs Step sessions, Rigidbody mid-solver values, preview budget scaling, and raw-capture tuple usageTest plan
dist/darwin-arm64/uloop skills install --claude --agentsregenerated copies; source and generated files verified byte-identical (including the two new references)references/links verified against existing targets