claude-dev-env 2.16.0 → 2.17.0

package/.agents/agents/AGENTS.md

@@ -2,6 +2,12 @@
 
 Agent definition files installed into `~/.agents/agents/` by `bin/install.mjs`. `~/.claude/agents` is a directory pointer to that folder. Each `.md` file defines a named subagent: its description (shown in the Claude Code UI), allowed tools, and behavioral instructions.
 
+Agent definitions use `~/.claude/rules/asd-ste100-language.md` for user-facing word choice, sentence style, tone, punctuation, and prose form. Each agent keeps its role-specific return format.
+
+When the global presentation envelope applies, ELI5 owns beginner framing,
+large visuals, minimal text, one stable self-contained HTML artifact,
+update-in-place continuity, and sharing.
+
 ## Agent files
 
 | File | Agent name | Role |

package/.agents/skills/AGENTS.md

@@ -2,6 +2,8 @@
 
 Each skill is a self-contained folder Claude Code loads on demand. At startup, only the skill's `name` and `description` metadata load. The full `SKILL.md` body and any support files load only when a skill becomes relevant to the conversation.
 
+Skill bodies use `~/.claude/rules/asd-ste100-language.md` for user-facing word choice, sentence style, tone, punctuation, and prose form. Each skill keeps its capability-specific output contract.
+
 ## Skill folder convention
 
 | Item | Role |
@@ -15,6 +17,8 @@ Each skill is a self-contained folder Claude Code loads on demand. At startup, o
 
 Skills install to `~/.agents/skills/<skill-name>/` via `packages/claude-dev-env/bin/install.mjs`. Claude Code looks them up at `~/.claude/skills/<skill-name>/`, a directory pointer to that folder. See `docs/references/skill-install-system.md` for the install pipeline.
 
+Retired skills live in `../skills-archived/` — versioned in the repo, not installed, not under the `.claude/skills` pointer. See `../skills-archived/AGENTS.md` for the archive index.
+
 ## Shared support code
 
 **`skills/_shared/`** — skill-local PR-loop helpers plus `@` stubs that name
@@ -24,49 +28,44 @@ contracts, runtime scripts). Map: `skills/_shared/.claude/CLAUDE.md`. End-of-run
 ## Skill groups
 
 **Planning and implementation**
-- `anthropic-plan` — creates a source-grounded plan packet before any code changes
-- `orchestrator` — turns the session into the orchestrator: it spawns executor subagents to do the code edits and test runs; hard decisions go to a shared advisor (Claude warm `session-advisor` via SendMessage; a third-party host: max-tier Claude via CLI Claude-chain)
-- `orchestrator-refresh` — sub-skill fired by the `/orchestrator` loop to re-assert the host-matched shared-advisor discipline mid-run (Claude SendMessage; a third-party host's Claude CLI chain, no Agent-tool advisor spawn)
-- `team-advisor` — binds one advisor at the strongest reachable tier (Claude warm agent; a third-party host: max-tier Claude via CLI Claude-chain, fail closed when unreachable); consult cadence and weight live in `docs/references/advisor-tool.md`
-- `grokify` — builds a paste-ready Grok Build handoff with a Claude advisor charter
-- `grok-spawn` — orchestrator playbook for fleets of headless grok CLI workers (preflight, batch spec, `spawn_grok_batch.py`)
+- `orchestrator` — turns the session into the orchestrator: it spawns executor subagents to do the code edits and test runs; hard decisions go to a shared advisor
+- `orchestrator-refresh` — re-asserts orchestrator discipline on a one-shot delayed wake
+- `team-advisor` — binds one advisor at the strongest reachable tier
+- `grok-spawn` — orchestrator playbook for fleets of headless grok CLI workers
 
 **PR review and convergence**
+
+- `shared-extraction-audit` — audits workflow packages for helpers that belong in shared libraries; extracts in small tested CLs
+- `name-by-capability-audit` — audits PR paths/titles for driver/motive words on reusable capability code
+- `review-tier` — classifies `review_tier_constants` from change axes, hard triggers, and user overrides
+- `review-router` — resolves and arms one supported `route_review_config` route through the registered Agent|Task gate
+- `pr-cleanup` — one-agent end-to-end PR cleanup: extraction → naming → sr-loop → small-cl; apply and validate fixes as they return
 - `autoconverge` — autonomous single-run workflow that drives a PR to ready
-- `pr-converge` — paced convergence loop across `ScheduleWakeup` ticks
-- `bugteam` — open-loop audit-fix until convergence
-- `copilot-review` — requests and polls a GitHub Copilot review
-- `copilot-finding-triage` — tiers each Copilot gate finding, verifies each code concern with an executed check, then routes it (auto-fix a confirmed defect, resolve a refuted one, page the user only for an inconclusive one)
-- `reviewer-gates` — availability gates for external reviewers (opt-out parse, Copilot quota, Bugbot trigger/detect)
-- `pr-fix-protocol` — applies reviewer findings as verified fixes and drives unresolved review threads to zero
-- `pr-loop-lifecycle` — opens and closes a PR-loop run (grant, teardown, PR description, revoke, report)
-- `pr-loop-cloud-transport` — six-step transport workflow that lets any PR-loop skill run in a session whose `gh` CLI is absent or cannot act on the PR (MCP schema load, origin/HEAD fix, identity rules, the gh-to-MCP substitution matrix, the Copilot status rule, and the post self-check)
-
-**Code review guides (Google eng-practices, adapted for an AI reviewer)**
-- `small-cl` — the Small CLs guide; refs: handling reviewer comments
-- `comments` — writing review comments and handling reviewer comments; refs: the review standard, pushback, respectful code reviews
-- `reviews` — the reviewer guide: standard, what to look for, navigating a CL, speed, pushback; refs: CL descriptions, review comments, emergencies, style guides, the two guide indexes
-- `descriptions` — writing good CL descriptions
-- `emergencies` — what counts as an emergency change; refs: what to look for, speed
-- Focused pull request guidance: [`small-cl/SKILL.md`](small-cl/SKILL.md)
+- `e-code-review` — max-recall code review at a selectable effort level
+- `e-simplify` — cleanup-only pass on the current diff
+- `small-cl` — the Small CLs guide for scope and split decisions
 
 **Research and discovery**
-- `recall` — retrieves facts from memory files
-- `remember` — saves a decision, gotcha, or architectural choice to the Obsidian vault
-- `everything-search` — file-system search via the Everything `es.exe` CLI
-- `beat-sheet` — reformats a reply as short single-line beats
-- `show` — creates and reviews inline visual explanations: diagrams, interactive widgets, mockups, charts, and illustrations
+- `everything-search` — file-system search via the Everything `es.exe` CLI on Windows
+- `eli5` — owns beginner framing, large visuals, minimal text, one stable self-contained HTML artifact, update-in-place continuity, and sharing
+
+**Source commands**
+- `source-command-logifix` — restore the Logitech Gaming Software tray icon on Windows
+- `source-command-sr-loop` — converging cleanup loop: simplify then code-review until clean
+
+**Samsung certification**
+- `cert-classification-rule` — add or change a Samsung cert-failure classification rule
 
 **Session and workflow management**
 - `session-log` — logs a session report to the Obsidian vault
 - `session-tidy` — tidies the session folder
 - `task-build` — gathers open tasks
-- `issue-tracker` — one consistent way to create, update in place, and close GitHub issues for a work-stream: one epic parent with native sub-issues, dedup-first, marker-delimited body sections edited in place, and an epic checklist mirroring the children
-- `closeout` — session-end entry that harvests obstacles into issue-candidate records and delegates filing to the `issue-tracker` agent (skill fallback), keeping the user-validation gate on each draft
-- `privacy-hygiene` — full-repo personal-data and secret sweep plus remediation guide
-- `update` — updates the dev-env package
-- `fresh-branch` — creates a clean branch off main
-- `rebase` — rebases onto main
-- `usage-pause` — waits out the 5-hour usage window in ScheduleWakeup stages that keep agent contexts warm; probes the OAuth usage endpoint or takes a manual reset override
+- `issue-tracker` — GitHub epic and sub-issue workflow
+- `privacy-hygiene` — full-repo personal-data and secret sweep
+- `update` — fast-forwards a local `main` branch
+- `fresh-branch` — creates a clean branch off main in a worktree
+- `rebase` — rebases onto main with verification gates
+- `usage-pause` — waits out the usage window in ScheduleWakeup stages
 - `skill-builder` — complete skill-building lifecycle
-- `auditing-claude-config` — audits a Claude Code setup for context-budget waste and produces a migration table with savings
+- `run-claude-dev-env` — build, install, and test the claude-dev-env installer
+- `prototype` — isolated hookless worktree sandbox for proof-of-concept builds

package/.agents/skills/_shared/pr-loop/AGENTS.md

@@ -1,16 +1,17 @@
 # pr-loop
 
-Shared infrastructure for the PR audit-fix loop used by `bugteam`,
-`pr-converge`, and `autoconverge`. Provides the XML prompt template, Python
-runtime scripts, the portable converge driver protocol, and named constants
-those skills invoke during each loop tick.
+Shared infrastructure for autoconverge PR convergence: XML prompt templates,
+portable-driver phase machine scripts under `skills/_shared/pr-loop/scripts/`,
+and runtime convergence/Codex scripts under `~/.claude/_shared/pr-loop/scripts/`.
+Retired entry skills (`bugteam`, `pr-converge`, and others) are archived under
+`packages/claude-dev-env/.agents/skills-archived/` for reference only.
 
 ## Subdirectories
 
 | Directory | Role |
 |---|---|
-| `prompts/` | XML agent prompt templates. |
-| `scripts/` | Python scripts for loop state management, prompt building, outcome recording, path resolution, pacer selection, and preflight checks. |
+| `prompts/` | XML agent prompt templates for internal lenses. |
+| `scripts/` | Portable pacer scripts: pacer selection, loop state, portable driver, prompt builders, handoff. |
 
 ## Canonical-path stubs (`~/.claude/_shared/pr-loop/`)
 
@@ -34,17 +35,23 @@ Open a stub, then load the `@` target:
 | File | Role |
 |---|---|
 | `portable-driver.md` | Continuous in-session pacer when Workflow / ScheduleWakeup are absent. |
-| `prompts/pr-consistency-audit.xml` | Structured prompt artifact for the cross-file consistency audit agent. |
+| `prompts/pr-consistency-audit.xml` | Cross-file consistency audit prompt for autoconverge self-review lanes. |
 | `scripts/select_converge_pacer.py` | Maps entry skill + host tool flags to `workflow`, `schedule_wakeup`, or `portable`. |
-| `scripts/build_audit_prompt.py` | Assembles the audit agent prompt from loop state and the constants module. |
-| `scripts/build_fix_prompt.py` | Assembles the fix agent prompt from loop state and findings. |
-| `scripts/init_loop_state.py` | Initializes the per-PR loop state JSON file. |
-| `scripts/write_audit_outcomes.py` | Writes the per-loop audit outcome XML into the workspace. |
-| `scripts/write_fix_outcomes.py` | Writes the per-loop fix outcome XML into the workspace. |
-| `scripts/preflight_worktree.py` | Verifies the working directory is a healthy worktree for the target PR's repo. |
+| `scripts/build_audit_prompt.py` | Assembles audit agent prompts from loop state (portable pacer). |
+| `scripts/build_fix_prompt.py` | Assembles fix agent prompts from loop state and findings. |
+| `scripts/init_loop_state.py` | Initializes per-PR loop state JSON for portable runs. |
+| `scripts/write_audit_outcomes.py` | Writes per-loop audit outcome XML into the workspace. |
+| `scripts/write_fix_outcomes.py` | Writes per-loop fix outcome XML into the workspace. |
+| `scripts/preflight_worktree.py` | Verifies the working directory is a healthy worktree for the target PR. |
 | `scripts/teardown_worktrees.py` | Removes loop worktrees on clean exit. |
-| `scripts/write_handoff.py` | Writes durable resume-handoff files under `~/.claude/runtime/pr-loop/<run-name>/` at each converge checkpoint. |
+| `scripts/write_handoff.py` | Writes durable resume-handoff files under `~/.claude/runtime/pr-loop/<run-name>/`. |
+| `scripts/portable_converge_driver.py` | Portable phase machine for autoconverge when `pacer=portable`. |
+| `scripts/build_converge_task_list.py` | Step-1 task list for portable runs. |
 | `scripts/_path_resolver.py` | Resolves workspace and worktree paths from PR metadata. |
 | `scripts/_cli_utils.py` | Shared CLI argument parsing helpers. |
 | `scripts/_xml_utils.py` | XML serialization helpers. |
-| `scripts/skills_pr_loop_constants/` | Named constants package imported by the scripts above. |
+| `scripts/skills_pr_loop_constants/` | Named constants for portable-driver scripts. |
+
+Runtime convergence checks, Codex review, and shared gates live in
+`packages/claude-dev-env/_shared/pr-loop/scripts/` (installed to
+`~/.claude/_shared/pr-loop/scripts/`).

package/.agents/skills/_shared/pr-loop/portable-driver.md

@@ -5,7 +5,7 @@ clean stamps, ready decisions, task lists, and “what next” never live as pro
 for the agent to invent. The agent runs scripts, reads JSON, and only performs
 judgment steps the JSON names.
 
-## Step 1 — task list (every autoconverge / portable run)
+## Step 1 — task list (every autoconverge portable run)
 
 ```
 python "$HOME/.claude/skills/_shared/pr-loop/scripts/build_converge_task_list.py" \
@@ -25,7 +25,7 @@ shared HEAD. Do not invent tasks in prose.
 
 ```
 python "$HOME/.claude/skills/_shared/pr-loop/scripts/select_converge_pacer.py" \
-  --skill <pr-converge|autoconverge> \
+  --skill autoconverge \
   --has-workflow <0|1> \
   --has-schedule-wakeup <0|1>
 ```
@@ -46,8 +46,7 @@ ScheduleWakeup path. When `pacer=portable`, use the control script below.
 3. Confirm the working directory is the PR’s own repo on the PR head SHA:
    `python "$HOME/.claude/skills/_shared/pr-loop/scripts/preflight_worktree.py" --owner <O> --repo <R> --mode strict`.
    Non-zero exit → report the `ABORT` line and stop.
-4. Cross-repo routing follows pr-converge Step 1.5: every local review and edit
-   runs with cwd set to the **PR worktree**.
+4. Cross-repo routing: every local review and edit runs with cwd set to the **PR worktree** (see autoconverge pre-flight and `preflight_worktree.py --mode strict`).
 
 `open-run` runs the same strict preflight before seeding state.
 
@@ -65,7 +64,7 @@ ok; `1` = contract failure; `2` = usage error.
 |---|---|
 | `open-run` | Require `portable`; preflight; seed state + task list; when `--codex-down` is off, set `codex_required` from CLI force-on or the weekly usage probe (same rule as `check_convergence`); next=`run_code_review` |
 | `after-code-review` | From returncode / dirty_tree / served_command |
-| `after-bugteam` | From pushed / converged |
+| `after-bugteam` | From pushed / converged (bug-audit phase on portable pacer) |
 | `after-bugbot` | From classification / inline lag |
 | `after-codex` | From classification clean / dirty / down |
 | `after-copilot-wait` | From review surfaced / wait cap |
@@ -105,7 +104,7 @@ Copilot quota pre-check:
 |---|---|---|
 | `run_code_review` | Run `commands` | `after-code-review` |
 | `apply_fixes_and_push` | Fix protocol; commit; push | re-review then `after-code-review` |
-| `run_bugteam` | resolve_worker_spawn / bugteam body | `after-bugteam` |
+| `run_bugteam` | Bug-audit lens / worker spawn body | `after-bugteam` |
 | `run_bugbot_gate` | Bugbot helper scripts | `after-bugbot` |
 | `run_codex_review` | Run Codex review step | `after-codex` |
 | `request_copilot_review` | Request Copilot | `after-copilot-wait` |
@@ -140,7 +139,7 @@ sections unchanged.
 | Worktree | `preflight_worktree.py` |
 | Code review | `$HOME/.claude/scripts/invoke_code_review.py` |
 | Workers | `$HOME/.claude/scripts/resolve_worker_spawn.py` |
-| Ready | `pr-converge/scripts/check_convergence.py` |
+| Ready | `$HOME/.claude/_shared/pr-loop/scripts/check_convergence.py` |
 | Handoff | `write_handoff.py` |
 
 ## Fail closed

package/.agents/skills/_shared/pr-loop/prompts/AGENTS.md

@@ -1,9 +1,9 @@
 # prompts
 
-XML agent prompt templates for the PR audit-fix loop.
+XML agent prompt templates for autoconverge internal lenses and portable pacer audits.
 
 ## Key files
 
 | File | Role |
 |---|---|
-| `pr-consistency-audit.xml` | Structured prompt artifact for the cross-file consistency audit agent. Defines the agent role, scope anchors, a ten-rule detection workflow, and output format. `bugteam` and `pr-converge` both inject this prompt when running the consistency audit step. |
+| `pr-consistency-audit.xml` | Structured prompt for the cross-file consistency audit. Defines the agent role, scope anchors, a ten-rule detection workflow, and output format. Autoconverge self-review and portable bug-audit lanes inject this prompt from `~/.claude/skills/_shared/pr-loop/prompts/pr-consistency-audit.xml`. |

package/.agents/skills/_shared/pr-loop/scripts/AGENTS.md

@@ -1,6 +1,6 @@
 # scripts
 
-Python scripts that run the PR audit-fix loop at runtime. Both `bugteam` and `pr-converge` invoke these scripts during each loop tick.
+Python scripts for autoconverge portable pacer control and loop helpers. Runtime convergence checks, Codex review, and shared gates live in `~/.claude/_shared/pr-loop/scripts/`.
 
 ## Two script homes
 

package/.agents/skills/_shared/pr-loop/scripts/portable_converge_driver.py

@@ -122,8 +122,6 @@ from skills_pr_loop_constants.portable_driver_constants import (  # noqa: E402
     CLI_SESSION_MODEL_FLAG,
     CLI_STATE_DIR_FLAG,
     CLI_STATE_FILE_FLAG,
-    CODEX_REVIEW_SCRIPTS_DIRNAME,
-    CODEX_REVIEW_SKILL_DIRNAME,
     COMMAND_AFTER_BUGBOT,
     COMMAND_AFTER_BUGTEAM,
     COMMAND_AFTER_CODE_REVIEW,
@@ -140,7 +138,6 @@ from skills_pr_loop_constants.portable_driver_constants import (  # noqa: E402
     EXIT_USAGE_ERROR,
     INLINE_LAG_STREAK_CAP,
     INVOKE_CODE_REVIEW_RELATIVE_PATH,
-    PORTABLE_SCRIPTS_TO_SKILLS_PARENT_HOPS,
     NEXT_APPLY_FIXES,
     NEXT_CHECK_READY,
     NEXT_MARK_READY,
@@ -200,19 +197,31 @@ from skills_pr_loop_constants.preflight_constants import (  # noqa: E402
     REPO_ARG_FLAG,
 )
 
-_skills_directory_for_codex = Path(__file__).resolve().parent
-_remaining_codex_path_hops = PORTABLE_SCRIPTS_TO_SKILLS_PARENT_HOPS
-while _remaining_codex_path_hops > 0:
-    _skills_directory_for_codex = _skills_directory_for_codex.parent
-    _remaining_codex_path_hops -= 1
-_codex_scripts_directory = (
-    _skills_directory_for_codex
-    / CODEX_REVIEW_SKILL_DIRNAME
-    / CODEX_REVIEW_SCRIPTS_DIRNAME
-)
-_codex_scripts_text = str(_codex_scripts_directory)
-if _codex_scripts_text not in sys.path:
-    sys.path.insert(0, _codex_scripts_text)
+def _resolve_shared_pr_loop_scripts_directory() -> Path:
+    """Return the installed or package ``_shared/pr-loop/scripts`` directory."""
+    scripts_directory = Path(__file__).resolve().parent
+    skill_directory = scripts_directory.parent
+    shared_skill_directory = skill_directory.parent
+    skills_directory = shared_skill_directory.parent
+    skills_home_directory = skills_directory.parent
+    package_or_user_home = (
+        skills_home_directory.parent
+        if skills_home_directory.name == ".agents"
+        else skills_home_directory
+    )
+    shared_under_package = package_or_user_home / "_shared" / "pr-loop" / "scripts"
+    shared_under_claude = (
+        package_or_user_home / ".claude" / "_shared" / "pr-loop" / "scripts"
+    )
+    if shared_under_package.is_dir():
+        return shared_under_package
+    return shared_under_claude
+
+
+_shared_pr_loop_scripts_directory = _resolve_shared_pr_loop_scripts_directory()
+_shared_scripts_text = str(_shared_pr_loop_scripts_directory)
+if _shared_scripts_text not in sys.path:
+    sys.path.insert(0, _shared_scripts_text)
 
 from codex_review_scripts_constants.codex_usage_probe_constants import (  # noqa: E402
     USAGE_REPORT_KEY_PERCENT_LEFT,

package/.agents/skills/_shared/pr-loop/scripts/skills_pr_loop_constants/portable_driver_constants.py

@@ -169,13 +169,9 @@ DEFAULT_SESSION_MODEL: str = "third-party"
 
 INVOKE_CODE_REVIEW_RELATIVE_PATH: str = ".claude/scripts/invoke_code_review.py"
 CHECK_CONVERGENCE_RELATIVE_PATH: str = (
-    ".claude/skills/pr-converge/scripts/check_convergence.py"
+    ".claude/_shared/pr-loop/scripts/check_convergence.py"
 )
 
-CODEX_REVIEW_SKILL_DIRNAME: str = "codex-review"
-CODEX_REVIEW_SCRIPTS_DIRNAME: str = "scripts"
-PORTABLE_SCRIPTS_TO_SKILLS_PARENT_HOPS: int = 3
-
 EXIT_SUCCESS: int = 0
 EXIT_USAGE_ERROR: int = 2
 EXIT_CONTRACT_ERROR: int = 1

package/.agents/skills/autoconverge/reference/convergence-gates.md

@@ -0,0 +1,31 @@
+# Convergence gates — printed labels
+
+`check_convergence.py` in `packages/claude-dev-env/_shared/pr-loop/scripts/`
+re-derives every readiness condition from GitHub and prints one numbered
+PASS/FAIL line per gate. When Bugbot is down, the bugbot review-body gate is
+omitted from the numbered list.
+
+## Exact printed labels (script order)
+
+When every gate is required and passes:
+
+1. `bugbot_clean_at == current_head`
+2. `bugbot review body clean` (omitted when Bugbot is down)
+3. `bugteam_clean_at == current_head`
+4. `copilot_clean_at == current_head`
+5. `codex_clean_at == current_head`
+6. `zero unresolved bot threads`
+7. `PR is mergeable`
+8. `no pending requested reviews`
+
+Success ends with:
+
+```
+All pre-conditions met — PR is ready to mark ready.
+```
+
+Failure ends with:
+
+```
+One or more pre-conditions not met — do not mark ready.
+```

package/.agents/skills/autoconverge/reference/convergence.md

@@ -6,7 +6,7 @@ Before the first round, the workflow checks once whether the PR branch conflicts
 with `origin/main`. When GitHub reports a conflict (`mergeable` false or
 `mergeable_state` dirty), one `clean-coder` rebases the branch onto `origin/main`
 and resolves every conflict — the edit remains in the working tree, review and
-verification follow the [review guide](../../reviews/SKILL.md#review-workflow),
+verification follow the [precatch rubric](../../_shared/pr-loop/precatch-rubric.md),
 and the commit step force-pushes with lease. The bug checks then run on a
 conflict-free diff.
 
@@ -62,15 +62,14 @@ confirmation gates that are expected to return zero.
 3. Run three reading lenses in parallel on that HEAD, each over the full
    `origin/main...HEAD` diff. Each lens receives the preflight's changed-file
 list and diffstat and reads only the files it needs from that list rather than
-re-deriving the diff; each lens applies the [review guide](../../reviews/SKILL.md#review-judgment).
+re-deriving the diff; each lens applies the [precatch rubric](../../_shared/pr-loop/precatch-rubric.md#review-judgment).
    - **Code-review lens** — a correctness-focused review pass (`code-quality-agent`),
      report-only workflow agent — see runCodeReviewLens in workflow/converge.mjs for its configuration.
      The built-in `/code-review` command is a separate surface outside this
      workflow path. On `pacer=workflow`, this lens runs inside `converge.mjs`.
-     On `pacer=portable`, the continuous driver uses the pr-converge CODE_REVIEW
-     phase through `invoke_code_review.py` instead of `runCodeReviewLens`
-     ([`../../_shared/pr-loop/portable-driver.md`](../../_shared/pr-loop/portable-driver.md);
-     [`../../pr-converge/reference/per-tick.md`](../../pr-converge/reference/per-tick.md)).
+     On `pacer=portable`, the continuous driver uses the CODE_REVIEW phase
+     through `invoke_code_review.py` instead of `runCodeReviewLens`
+     ([`../../_shared/pr-loop/portable-driver.md`](../../_shared/pr-loop/portable-driver.md)).
    - **Bug-audit lens** — the bug-audit (`code-quality-agent`) applying the
      shared A–P rubric from `_shared/pr-loop/audit-contract.md`, then its
      adversarial second pass, and the doc-parity, test-assertion, and
@@ -85,7 +84,7 @@ re-deriving the diff; each lens applies the [review guide](../../reviews/SKILL.m
    text, and collects every distinct bot thread id so the fix lens resolves all
    colliding threads.
 5. **Any findings** → one `clean-coder` applies every fix per the
-   `pr-fix-protocol` skill (`../../pr-fix-protocol/SKILL.md`): a single
+   shared [fix protocol](../../_shared/pr-loop/fix-protocol.md): a single
    test-first commit, a push, then a reply and resolve on each finding that
    carries a GitHub review thread. Before its turn ends, the edit step dry-runs
    the CODE_RULES commit gate (`code_rules_gate.py --staged`) over its staged
@@ -195,7 +194,7 @@ bugbot-flagged path does not satisfy the round; its clean verdict is not a clean
 
 ## The ready definition
 
-`check_convergence.py` (`pr-converge/scripts/check_convergence.py`) is the single
+`check_convergence.py` (`packages/claude-dev-env/_shared/pr-loop/scripts/check_convergence.py`) is the single
 source of truth for readiness. Ready means the script exits `0` and prints:
 
 ```
@@ -204,8 +203,7 @@ All pre-conditions met — PR is ready to mark ready.
 
 The script re-derives every condition from GitHub and prints one PASS/FAIL line
 per label. The exact printed labels (script order) live in
-[`pr-converge/reference/convergence-gates.md`](../../pr-converge/reference/convergence-gates.md)
-§ (f) Mark ready and report — the seven-label block under "Exact printed labels".
+[`convergence-gates.md`](convergence-gates.md).
 
 The script has no Claude APPROVED review gate. Agent-side checks (Claude
 reviewer presence, broader unresolved-thread sweeps) sit outside this machine

package/.agents/skills/autoconverge/reference/gotchas.md

@@ -52,7 +52,7 @@ fails in a new way.
 
 - **Commit or push fails after a repair.** Read the hook or command output,
   repair the named issue, and record the rerun. Apply the
-  [review guide](../../reviews/SKILL.md#review-workflow) before retrying.
+  [precatch rubric](../../_shared/pr-loop/precatch-rubric.md) before retrying.
 
 - **A Copilot "down" verdict is valid only after the full poll budget.** A
   successful review request means the review is in flight — Copilot typically

package/.agents/skills/autoconverge/reference/stop-conditions.md

@@ -59,7 +59,7 @@ skill still runs teardown (revoke permissions, final report).
 
 ## Not a blocker (the run continues)
 
-Each reviewer-down condition below skips its own convergence-check gate. The flag and the matching `CLAUDE_REVIEWS_DISABLED` token for each condition are the single flag-per-condition table in [`../../reviewer-gates/SKILL.md`](../../reviewer-gates/SKILL.md) § "Convergence-check bypass flags"; the token carries the bypass into the mark-ready blocker hook's no-flag re-check.
+Each reviewer-down condition below skips its own convergence-check gate. The flag and the matching `CLAUDE_REVIEWS_DISABLED` token for each condition are the single flag-per-condition table in [`../../skills-archived/reviewer-gates/SKILL.md`](../../skills-archived/reviewer-gates/SKILL.md) — "Convergence-check bypass flags"; the token carries the bypass into the mark-ready blocker hook's no-flag re-check.
 
 - **Clean-audit post bypassed** — every review lens is clean on HEAD, but the
   environment refuses the CLEAN bugteam review post (the `post_audit_thread.py`

package/.agents/skills/autoconverge/workflow/converge.codex-gate.test.mjs

@@ -413,7 +413,7 @@ test('CONFIG expands home for the codex-review scripts directory', () => {
   );
   assert.match(
     convergeSource,
-    /codexScripts:\s*`\$\{homeDirectory\}\/\.claude\/skills\/codex-review\/scripts`/,
+    /codexScripts:\s*`\$\{homeDirectory\}\/\.claude\/_shared\/pr-loop\/scripts`/,
     'expected codexScripts to expand homeDirectory rather than embed $HOME',
   );
   assert.doesNotMatch(

package/.agents/skills/autoconverge/workflow/converge.contract.test.mjs

@@ -19,7 +19,7 @@ function loadReviewResultParser() {
 
 test('review tasks apply the canonical review guide and return a review result', () => {
   assert.match(workflowSource, /function runReviewTask\(task, context\)/)
-  assert.match(workflowSource, /skills\/reviews\/SKILL\.md#review-workflow/)
+  assert.match(workflowSource, /_shared\/pr-loop\/precatch-rubric\.md/)
   assert.match(workflowSource, /REVIEW_RESULT: PASS/)
   assert.match(workflowSource, /REVIEW_RESULT: FINDINGS/)
 })

package/.agents/skills/autoconverge/workflow/converge.mjs

@@ -40,15 +40,15 @@ const CONFIG = {
   codexReviewTimeoutSeconds: 2400,
   codexReviewPollIntervalSeconds: 120,
   codexReviewMaxPolls: 21,
-  sharedScripts: '$HOME/.claude/skills/pr-converge/scripts',
+  sharedScripts: '$HOME/.claude/_shared/pr-loop/scripts',
   prLoopScripts: '$HOME/.claude/_shared/pr-loop/scripts',
-  codexScripts: `${homeDirectory}/.claude/skills/codex-review/scripts`,
+  codexScripts: `${homeDirectory}/.claude/_shared/pr-loop/scripts`,
   bugteamRubric: '$HOME/.claude/_shared/pr-loop/audit-contract.md',
   precatchRubric: '$HOME/.claude/_shared/pr-loop/precatch-rubric.md',
 }
 
 const REVIEWER_GATE_SENTINEL = 'CLAUDE_REVIEWER_GATE=autoconverge '
-const REVIEW_GUIDE_REFERENCE = 'skills/reviews/SKILL.md#review-workflow'
+const REVIEW_GUIDE_REFERENCE = '_shared/pr-loop/precatch-rubric.md'
 const REVIEW_RESULT_PASS_MARKER = 'REVIEW_RESULT: PASS'
 const REVIEW_RESULT_FINDINGS_MARKER = 'REVIEW_RESULT: FINDINGS'
 const REVIEW_RESULT_INSTRUCTION =