Repository navigation
Conversation
Move the line-capture and line-analyze commands to .agents/skills, with their scripts next to them, and link .claude/skills to .agents/skills so Claude Code finds them. The capture launcher now runs a dependency preflight first and works on macOS, Linux and Windows (Git Bash). The CDP recorder releases every page-like target so extension windows no longer open frozen. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Writes the numbered test plan (test/<feature>/PLAN.md) for a bridge feature before any capture or code: lettered sections, stable step IDs, before/capture/after phases and a Result column, plus labeled sample media when the feature needs files. check-plan.py validates the layout and make-samples.py generates the samples. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Turns a feature's test screenshots into PR-ready evidence: collects them from local folders, builds a local contact sheet to pick from, finds names, faces, phone numbers, emails and MIDs with on-device OCR (Apple Vision on macOS, tesseract elsewhere), covers everything outside the test plan's test data with solid boxes, and writes approved, metadata-free PNGs plus a manifest. Raw screenshots never leave the machine. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sends test messages and files into LINE chats through Beeper Desktop's local API and reads back what Beeper received: type, mimetype, size, width/height, GIF flag, file hash and frame count. Saves before/after snapshots and diffs them, so tests no longer depend on sending from a phone and screenshotting Beeper. It only reads and sends in chats the user allowlists on first use; the token comes from the keychain or BEEPER_ACCESS_TOKEN and is never printed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
line-implement now carries the whole workflow that used to be split between the skill and the line-implement command: it starts from an optional line-test-plan, captures with line-capture, analyzes with line-analyze and can finish with line-evidence. The .agents/commands folder and its .claude/commands link are gone, since every command is now a skill. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
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.
Updates
line-implementto use the other LINE skills, and removes the old.agents/commandsfolder now that every command is a skill. Depends on #291: review that one first, and only the last commit here is new.🇬🇧 More details
Summary
line-implementis the skill that runs a whole feature: plan, capture, analyze, implement, verify. It now takes in the steps from the oldline-implementcommand, and it points to the other skills for each step.Why
Part 5 of 5 of the stack that brings the LINE skills into this repo, as asked on #280. Before, the workflow was split between
.agents/skills/line-implementand.agents/commands/line-implement.md, and the two had drifted apart.Changes
.agents/skills/line-implement/SKILL.mdandREFERENCE.md: the full workflow in one place.line-test-plan.line-captureandline-analyze, with their preflight and script paths.line-evidence..agents/commands/line-implement.mdis removed. It was the last file in.agents/commands/, so the.claude/commandslink is removed too.How it was tested
pre-commit runon every changed file: passed.Not tested (and why)
Related
🤖 Generated with Claude Code using clins:create-pr