@event4u/agent-config 1.9.1 → 1.13.0

package/.agent-src/rules/roadmap-progress-sync.md

@@ -1,6 +1,6 @@
 ---
 type: "auto"
-description: "Editing checkboxes in agents/roadmaps/*.md — [x], [~], [-], or add/rename/remove phases — must run task roadmap-progress in the SAME response; a roadmap that hits 0 open items must also be archived in the SAME response"
+description: "Editing checkboxes in agents/roadmaps/*.md — [x], [~], [-], or add/rename/remove phases — must regenerate the roadmap dashboard in the SAME response; a roadmap that hits 0 open items must also be archived in the SAME response"
 alwaysApply: false
 source: package
 ---
@@ -11,7 +11,7 @@ source: package
 
 **CRITICAL — ZERO TOLERANCE:** Whenever you change checkbox state in a
 roadmap file (`agents/roadmaps/*.md`, module or package equivalents)
-you MUST run `task roadmap-progress` **in the same response** — not
+you MUST regenerate the dashboard **in the same response** — not
 later, not batched across sessions, not "at the end of the roadmap".
 
 `agents/roadmaps-progress.md` is the read-only dashboard. Every
@@ -24,33 +24,44 @@ roadmap to `count_open == 0` (every item is `[x]`, `[~]`, or `[-]`),
 in `agents/roadmaps/` is a rule violation. See `roadmap-management`
 for the archive vs skipped decision table.
 
+## How to regenerate
+
+```bash
+./agent-config roadmap:progress
+```
+
+The `./agent-config` wrapper is written into the project root by the
+installer and delegates to the master CLI inside
+`node_modules/@event4u/agent-config/` or `vendor/event4u/agent-config/`.
+No global tooling required.
+
 ## Triggers
 
 | Edit | Must run, same response |
 |---|---|
-| Mark step `[x]`, `[~]`, `[-]`, or unmark back to `[ ]` | `task roadmap-progress` |
-| Add, rename, or remove a phase | `task roadmap-progress` |
-| Create a new roadmap file | `task roadmap-progress` |
-| **Last `[ ]` flips** — roadmap reaches `count_open == 0` | `git mv` → `archive/` (or `skipped/`) **then** `task roadmap-progress` |
-| Move roadmap between `roadmaps/` ↔ `archive/` ↔ `skipped/` | `task roadmap-progress` |
+| Mark step `[x]`, `[~]`, `[-]`, or unmark back to `[ ]` | regenerate dashboard |
+| Add, rename, or remove a phase | regenerate dashboard |
+| Create a new roadmap file | regenerate dashboard |
+| **Last `[ ]` flips** — roadmap reaches `count_open == 0` | `git mv` → `archive/` (or `skipped/`) **then** regenerate dashboard |
+| Move roadmap between `roadmaps/` ↔ `archive/` ↔ `skipped/` | regenerate dashboard |
 
 **Batching:** multiple checkbox edits in one response → a **single**
-`task roadmap-progress` call at the end is enough. If one closes a
-roadmap, archive it first, then run the single regen. But the
-response must not end without it.
+regeneration at the end is enough. If one closes a roadmap, archive
+it first, then run the single regen. But the response must not end
+without it.
 
-## Why a rule, not just a skill tip
+## Why this is a rule, not a skill tip
 
 The `roadmap-management` skill documents the command in several
 places, but skill body text is easy to miss under procedure pressure.
 A rule collapses the constraint into one line the model cannot skip:
-"checkbox edit → `task roadmap-progress` — same response".
+"checkbox edit → regenerate dashboard — same response".
 
 ## Do NOT
 
 - Do NOT edit `agents/roadmaps-progress.md` by hand — always regenerate.
 - Do NOT defer regen to "next commit" or "before push" — same response.
-- Do NOT rely on `task ci` / `task roadmap-progress-check` as first line of defence — CI is last-line, not real-time.
+- Do NOT rely on CI (`--check` mode) as first line of defence — CI is last-line, not real-time.
 - Do NOT skip regen because "only one checkbox changed" — the dashboard aggregates counts and phase percentages that shift on single edits.
 - Do NOT leave a 100%-complete roadmap in `agents/roadmaps/` "for review" — archive same response, ask the user afterwards if needed, not before.
 - Do NOT regenerate the dashboard before the `git mv` when a roadmap closes — otherwise it reappears in "Open roadmaps".

package/.agent-src/rules/size-enforcement.md

@@ -26,4 +26,4 @@ source: package
 
 → Frontmatter contract (required/optional keys per type):
 [`agents/docs/frontmatter-contract.md`](../../../agents/docs/frontmatter-contract.md).
-Schemas live in `scripts/schemas/` and are enforced by `task validate-schema`.
+Schemas live in `scripts/schemas/` and are enforced by `python3 scripts/validate_frontmatter.py`.

package/.agent-src/rules/skill-quality.md

@@ -28,7 +28,7 @@ Every skill's YAML frontmatter MUST validate against `scripts/schemas/skill.sche
 See [`agents/docs/frontmatter-contract.md`](../../../agents/docs/frontmatter-contract.md)
 for the human-readable contract across all artefact types. Violations are
 reported by `scripts/skill_linter.py` as `schema_<rule>` errors and fail
-`task validate-schema` / `task ci`.
+`python3 scripts/validate_frontmatter.py` and the full CI pipeline.
 
 ## Description Triggering
 

package/.agent-src/rules/think-before-action.md

@@ -16,6 +16,7 @@ source: package
 - If requirements are unclear, ask a precise clarification question instead of making hidden assumptions
 - Refactors must preserve behavior, validation, examples, and anti-failure guidance unless there is an explicit reason to change them
 - Do NOT modify code you do not fully understand — read it first, trace the flow, then change it
+- Multiple valid frameworks/patterns already in the codebase (Tailwind + Flux, multiple form libs, competing state stores) → do NOT pick one silently, ask. See [`no blind implementation`](../guidelines/agent-infra/agent-interaction-and-decision-quality.md#2-no-blind-implementation)
 
 ## The Developer Workflow
 

package/.agent-src/rules/user-interaction.md

@@ -25,30 +25,76 @@ The user should be able to reply with just a number (e.g., `1`) instead of typin
 - **Every question with choices** must use numbered options — no exceptions.
 - **Keep options short** — one line each, with a brief explanation after the dash.
 - **Always include a "skip" or "no change" option** when applicable.
-- **Default/recommended option** can be marked: `1. Do X (recommended)`.
+- **Always state a recommendation** — see iron law below.
 - **Use the user's language** for the question and options.
 - **Accept both** the number and a natural language answer (e.g., "1" or "the first one").
 
+### Iron Law — ALWAYS recommend
+
+```
+EVERY numbered-option question MUST state which option the agent recommends and WHY.
+"Egal, was bevorzugst Du?" / "no preference" is NEVER an acceptable agent stance.
+```
+
+The user is asking the agent because the agent has read the code, the
+contracts, the trade-offs. Refusing to take a position dumps that work
+back on the user. Take the position; be wrong out loud if needed.
+
+**Format:**
+
+- Mark the recommended option inline: `1. Do X — short explanation (recommended)`.
+- After the option block, state **WHY** in 1–3 sentences: the trade-off
+  that tips the balance, plus the **caveat** that would flip it.
+- If the agent genuinely cannot pick (rare — true 50/50 with missing
+  data), say what data would break the tie and ask for that instead.
+
+**Example:**
+
+```
+> 1. Hybrid contract — keys + query (recommended)
+> 2. Key-based — extend the package
+> 3. Semantic — change all call sites
+
+I recommend 1: solves the acute consult-flow without a cross-repo PR,
+and the file-fallback stays trivial. Caveat — if hit-rate on the
+concat-shim turns out poor in practice, escalate to 2.
+```
+
+**What does NOT count as a recommendation:**
+
+- "Both work" / "either is fine" / "depends on what you prefer"
+- Listing pros and cons without picking
+- "I'd lean towards X" without a reason
+- Hiding behind "you know the project better" (the agent just researched it)
+
 ### Examples
 
-**Binary choice:**
+**Binary choice (with recommendation):**
 ```
-> 1. Interactive — ask before each comment
+> 1. Interactive — ask before each comment (recommended)
 > 2. Automatic — handle all independently
+
+I recommend 1: the comments touch security-sensitive code, so a wrong
+auto-fix is more expensive than the friction of approving each one.
+Switch to 2 if the comments turn out to be pure formatting.
 ```
 
-**Multiple choice with skip:**
+**Multiple choice with skip (with recommendation):**
 ```
-> 1. Fix the code
+> 1. Fix the code (recommended)
 > 2. Fix the test
 > 3. Skip
+
+I recommend 1: the test is asserting the documented behavior; the
+production code drifted from the contract. Pick 2 only if the contract
+itself is wrong.
 ```
 
-**Confirmation with context:**
+**Confirmation with context (recommendation implicit in framing):**
 ```
 > Found PR #1399 on branch `chore/refactor-agent-setup-2`.
 >
-> 1. Yes, that's the right PR
+> 1. Yes, that's the right PR (recommended — branch matches)
 > 2. No, different PR — I'll provide the URL
 ```
 

package/.agent-src/scripts/update_roadmap_progress.py

@@ -19,6 +19,12 @@ Percentage = done / (done + open). Deferred and cancelled do not count towards
 Invocation (from project root):
   python3 .augment/scripts/update_roadmap_progress.py              # rewrite
   python3 .augment/scripts/update_roadmap_progress.py --check      # CI: exit 1 if stale
+
+`--check` mode also fails when a roadmap reaches `count_open == 0` but is
+still under `agents/roadmaps/` instead of `agents/roadmaps/archive/` —
+backstopping the `roadmap-progress-sync` rule's "completion = archival,
+same response" requirement. The write path emits the same finding as a
+warning on stderr and still regenerates the dashboard.
 """
 
 from __future__ import annotations
@@ -30,9 +36,21 @@ from dataclasses import dataclass, field
 from pathlib import Path
 
 CHECKBOX_RE = re.compile(r"^\s*[-*]\s+\[([ xX~\-])\]\s", re.MULTILINE)
-# H2 or H3 heading starting with "Phase <number>"; separator (colon, em-dash,
-# hyphen, or whitespace) and name are optional.
-PHASE_RE = re.compile(r"^(#{2,3})\s+Phase\s+(\d+)(?:[\s:\u2014\-]+(.*?))?\s*$", re.MULTILINE)
+# H2 or H3 heading starting with "Phase <id>"; separator (colon, em-dash,
+# hyphen, or whitespace) and name are optional. The id supports three
+# project-level conventions:
+#   - numeric        `Phase 0`, `Phase 10`
+#   - roman I..XXXIX `Phase I`, `Phase III`
+#   - letter track   `Phase A`, `Phase B1` (single uppercase letter,
+#                    optional trailing digits for sub-track IDs)
+# Roman is capped at [IVX]+ (up to XXXIX) on purpose: the broader
+# [IVXLCDM]+ would also match all-caps words like `Phase LIVE`. Letter
+# is [A-Z] not [A-Za-z] so `## Phase overview` stays a non-phase anchor.
+PHASE_RE = re.compile(
+    r"^(#{2,3})\s+Phase\s+(\d+|[IVX]+|[A-Z](?:\d+)?)"
+    r"(?:[\s:\u2014\-]+(.*?))?\s*$",
+    re.MULTILINE,
+)
 TITLE_RE = re.compile(r"^#\s+(?:Roadmap:\s*)?(.+?)\s*$", re.MULTILINE)
 EXCLUDE_NAMES = {"template.md", "README.md", "progress.md", "roadmaps-progress.md"}
 EXCLUDE_PREFIXES = ("open-questions",)
@@ -41,7 +59,10 @@ EXCLUDE_DIRS = {"archive", "skipped"}
 
 @dataclass
 class PhaseStats:
-    number: int
+    # Phase identifier as it appears in the heading: numeric ("0"),
+    # roman ("III"), or letter-track ("A", "B1"). Kept as a string so
+    # non-numeric conventions survive round-tripping through render().
+    id: str
     name: str
     done: int = 0
     open_: int = 0
@@ -139,9 +160,9 @@ def parse_roadmap(path: Path, roadmap_root: Path) -> RoadmapStats | None:
         start = pm.end()
         end = phase_matches[i + 1].start() if i + 1 < len(phase_matches) else len(text)
         d, o, df, c = count_checkboxes(text[start:end])
-        number = int(pm.group(2))
-        name = (pm.group(3) or "").strip() or f"Phase {number}"
-        stats.phases.append(PhaseStats(number, name, d, o, df, c))
+        phase_id = pm.group(2)
+        name = (pm.group(3) or "").strip() or f"Phase {phase_id}"
+        stats.phases.append(PhaseStats(phase_id, name, d, o, df, c))
     return stats
 
 
@@ -161,6 +182,15 @@ def collect(roadmap_root: Path) -> list[RoadmapStats]:
     return results
 
 
+def unarchived_complete(roadmaps: list[RoadmapStats]) -> list[RoadmapStats]:
+    # A roadmap is complete when every active checkbox is done and at least
+    # one active checkbox exists. The `roadmap-progress-sync` rule mandates
+    # that such a roadmap be moved to `agents/roadmaps/archive/` in the
+    # same response that closes its last open item; `collect()` already
+    # excludes that directory, so anything left here is unarchived.
+    return [r for r in roadmaps if r.total_active > 0 and r.open_ == 0]
+
+
 def render(roadmaps: list[RoadmapStats]) -> str:
     total_done = sum(r.done for r in roadmaps)
     total_active = sum(r.total_active for r in roadmaps)
@@ -201,7 +231,7 @@ def render(roadmaps: list[RoadmapStats]) -> str:
         lines.append("|---|---|---|---:|---:|---:|---:|---:|")
         for p in r.phases:
             lines.append(
-                f"| {p.number} | {p.name} | {p.state} | {p.done} | {p.open_} | "
+                f"| {p.id} | {p.name} | {p.state} | {p.done} | {p.open_} | "
                 f"{p.deferred} | {p.cancelled} | {p.percent}% |"
             )
         lines.append("")
@@ -226,10 +256,22 @@ def main() -> int:
     roadmaps = collect(roadmap_root)
     new_text = render(roadmaps)
     current = target.read_text(encoding="utf-8") if target.exists() else ""
+    complete = unarchived_complete(roadmaps)
     if args.check:
-        if current != new_text:
+        stale = current != new_text
+        if stale:
             print(f"❌  {target.relative_to(args.repo_root)} is stale. "
-                  f"Run `task roadmap-progress` to regenerate.", file=sys.stderr)
+                  f"Run `python3 .augment/scripts/update_roadmap_progress.py` "
+                  f"to regenerate (or `task roadmap-progress` in Taskfile "
+                  f"projects).", file=sys.stderr)
+        if complete:
+            print("❌  Completed roadmaps are still in `agents/roadmaps/` — "
+                  "move them to `agents/roadmaps/archive/` (per the "
+                  "`roadmap-progress-sync` rule):", file=sys.stderr)
+            for r in complete:
+                print(f"      - {r.rel}  ({r.done}/{r.total_active} done)",
+                      file=sys.stderr)
+        if stale or complete:
             return 1
         print(f"✅  {target.relative_to(args.repo_root)} is up to date.")
         return 0
@@ -237,6 +279,11 @@ def main() -> int:
     print(f"✅  Wrote {target.relative_to(args.repo_root)} · "
           f"{len(roadmaps)} roadmap(s) · "
           f"{sum(r.done for r in roadmaps)}/{sum(r.total_active for r in roadmaps)} steps done.")
+    if complete:
+        print("⚠️   Completed roadmaps not yet archived — move to "
+              "`agents/roadmaps/archive/`:", file=sys.stderr)
+        for r in complete:
+            print(f"      - {r.rel}", file=sys.stderr)
     return 0
 
 

package/.agent-src/skills/check-refs/SKILL.md

@@ -22,7 +22,7 @@ execution:
 - CI `check-refs` job failed and target ref needs locating
 
 NOT: body-only edits, frontmatter shape (use `lint-skills`), compressed pairs
-(use `task sync-check`).
+(use `bash scripts/compress.sh --check`).
 
 ## Procedure
 

package/.agent-src/skills/command-routing/SKILL.md

@@ -56,7 +56,7 @@ Read `github.pr_reply_method` from `.agent-settings.yml`:
 
 ### 2. Bot icon prefix
 
-Read `project.pr_comment_bot_icon` from `.agent-settings.yml`:
+Read `personal.pr_comment_bot_icon` from `.agent-settings.yml`:
 - `true` → prefix reply body with `🤖 `.
 - `false` or not set → no prefix.
 

package/.agent-src/skills/command-writing/SKILL.md

@@ -102,10 +102,11 @@ multi-paragraph explanation, extract it into a skill and call it.
 
 * Run `python3 scripts/skill_linter.py .agent-src.uncompressed/commands/{name}.md`
   → 0 FAIL.
-* Run `task sync` → regenerates `.agent-src/commands/{name}.md`.
-* Run `task generate-tools` → creates the Claude symlink at
+* Run `bash scripts/compress.sh --sync` → regenerates `.agent-src/commands/{name}.md`.
+* Run `python3 scripts/compress.py --generate-tools` → creates the Claude symlink at
   `.claude/skills/{name}/SKILL.md`.
-* Run `task ci` → exits 0 except for tolerated warnings.
+* Run the full CI pipeline locally — each script directly — must exit 0
+  except for tolerated warnings.
 
 ## Output format
 

package/.agent-src/skills/file-editor/SKILL.md

@@ -50,7 +50,7 @@ At the start of a conversation (or on first file edit), read `.agent-settings.ym
 cat .agent-settings.yml 2>/dev/null
 ```
 
-- If the file doesn't exist → **do not open files**. Suggest running `/config-agent-settings`.
+- If the file doesn't exist → **do not open files**. Suggest running `scripts/install` (then `/onboard`).
 - If `personal.open_edited_files: false` → do nothing.
 - If `personal.open_edited_files: true` and `personal.ide` is set → open files after edits.
 
@@ -124,6 +124,6 @@ code app/Models/User.php
 ## Do NOT
 
 - Do NOT open files without checking `.agent-settings.yml` first.
-- Do NOT prompt the user about IDE settings during normal work — suggest `/config-agent-settings` instead.
+- Do NOT prompt the user about IDE settings during normal work — suggest `/onboard` (for first-run) or editing `.agent-settings.yml` directly.
 - Do NOT open files that were only read, not edited.
 - Do NOT open more than 10 files at once — summarize instead.

package/.agent-src/skills/guideline-writing/SKILL.md

@@ -107,16 +107,17 @@ Above the split signal, break by sub-topic into sibling files in the same folder
 
 * Run `python3 scripts/skill_linter.py .agent-src.uncompressed/guidelines/{topic}/{name}.md`
   → 0 FAIL (guidelines have relaxed linting but must still parse).
-* Run `task sync` → regenerates `.agent-src/guidelines/`.
+* Run `bash scripts/compress.sh --sync` → regenerates `.agent-src/guidelines/`.
 * Run `python3 scripts/check_references.py` → no broken links.
-* Run `task ci` → exits 0 except for tolerated warnings.
+* Run the full CI pipeline locally — each script directly — must exit 0
+  except for tolerated warnings.
 
 ## Output format
 
 1. Complete guideline at `.agent-src.uncompressed/guidelines/{topic}/{name}.md`
 2. At least one skill or rule linking to it
 3. Linter + `check_references.py` clean
-4. `task sync` confirmation
+4. `bash scripts/compress.sh --sync` confirmation
 
 ## Gotchas
 

package/.agent-src/skills/learning-to-rule-or-skill/SKILL.md

@@ -169,8 +169,8 @@ Mandatory fields the draft MUST fill:
 * `Success signal` (§7) — one metric, one baseline, one target, one
   evaluation date
 
-Run `python3 scripts/check_proposal.py agents/proposals/<id>.md` (or
-`task check-proposal`) before handing to `upstream-contribute`. The
+Run `python3 scripts/check_proposal.py agents/proposals/<id>.md` before
+handing to `upstream-contribute`. The
 script is a hard gate: non-zero exit = the proposal does not move
 to stage `gated`.
 

package/.agent-src/skills/lint-skills/SKILL.md

@@ -23,7 +23,7 @@ execution:
 - Investigate reported linter failure
 
 NOT: single file (call `skill_linter.py <path>`), cross-refs (use `check-refs`),
-compression freshness (use `task sync-check`).
+compression freshness (use `bash scripts/compress.sh --check`).
 
 ## Procedure
 

package/.agent-src/skills/roadmap-management/SKILL.md

@@ -23,8 +23,8 @@ Do NOT use when:
 
 `agents/roadmaps-progress.md` is auto-generated and must reflect the
 live state in real time. After **any** checkbox edit (`[x]`, `[~]`,
-`[-]`, `[ ]`) or phase add/rename/remove in a roadmap file, run
-`task roadmap-progress` **in the same response**.
+`[-]`, `[ ]`) or phase add/rename/remove in a roadmap file, regenerate
+the dashboard **in the same response**.
 
 **Completion = archival.** If an edit takes a roadmap to
 `count_open == 0` (pure `[x]`, or `[x]` + `[~]`/`[-]`), `git mv`
@@ -34,14 +34,14 @@ below. A 100%-complete roadmap left in `agents/roadmaps/` makes
 the next reader think work is still open.
 
 Enforced by [`roadmap-progress-sync`](../../rules/roadmap-progress-sync.md).
-Batching edits in one response is fine — one final `task roadmap-progress`
-before replying is enough. But the response must not end without it.
+Batching edits in one response is fine — one final regeneration before
+replying is enough. But the response must not end without it.
 
 ## Procedure: Manage a roadmap
 
 1. **Identify need** — Is this a multi-step change that spans sessions or agents?
 2. **Create or locate** — Create new roadmap in `agents/roadmaps/` or find existing one.
-3. **Update progress** — Mark completed steps with `[x]`, add notes for blockers, then run `task roadmap-progress` in the same response (enforced by `roadmap-progress-sync`).
+3. **Update progress** — Mark completed steps with `[x]`, add notes for blockers, then regenerate the dashboard in the same response (enforced by `roadmap-progress-sync`).
 4. **Verify** — Confirm all steps reflect current state, no stale information.
 
 A roadmap is a structured `.md` file in `agents/roadmaps/` that describes a multi-step change
@@ -139,7 +139,7 @@ Every roadmap implicitly includes these gates (run after each step that changes
 2. Use the template structure from `.augment/templates/roadmaps.md`.
 3. Review with the user iteratively until approved.
 4. Save with a kebab-case filename (e.g. `optimize-webhook-jobs.md`).
-5. Run `task roadmap-progress` so the dashboard includes the new roadmap.
+5. Regenerate the dashboard so the new roadmap is included.
 
 ### Executing a roadmap
 
@@ -147,7 +147,7 @@ Every roadmap implicitly includes these gates (run after each step that changes
 2. Find the next unchecked step (`- [ ]`).
 3. Summarize what needs to be done.
 4. Ask the user before implementing (numbered options: implement / adjust / skip).
-5. After implementation: mark `[x]`, run quality gates, then `task roadmap-progress`.
+5. After implementation: mark `[x]`, run quality gates, then regenerate the dashboard.
 6. Move to the next step.
 
 ### Resuming a roadmap
@@ -228,7 +228,7 @@ After the last step of a roadmap is done, check completion status:
    git mv agents/roadmaps/{file} agents/roadmaps/skipped/{file}
    ```
 
-6. **Regenerate the dashboard:** `task roadmap-progress`. The moved roadmap is
+6. **Regenerate the dashboard** (see "Command" below). The moved roadmap is
    excluded from the active set once it sits in `archive/` or `skipped/`.
 
 ### When to use `skipped/` vs `archive/`
@@ -262,10 +262,13 @@ is rewritten by `.augment/scripts/update_roadmap_progress.py`.
 Command:
 
 ```bash
-task roadmap-progress          # rewrite the dashboard
-task roadmap-progress-check    # CI: fail if stale
+./agent-config roadmap:progress           # rewrite the dashboard
+./agent-config roadmap:progress-check     # CI: fail if stale
 ```
 
+The `./agent-config` wrapper is written into the project root by the
+installer and always works — no global tooling or task runner required.
+
 The dashboard is a **read-only snapshot**. Do not edit it by hand — regenerate it.
 
 ## Output format

package/.agent-src/skills/rtk-output-filtering/SKILL.md

@@ -38,43 +38,33 @@ rtk docker compose ps # compact container status
 
 ## Procedure: Wrap commands with rtk
 
-1. Run `which rtk` (silent) — check if installed.
-2. **If installed** → prefix commands with `rtk` when output >30 lines expected.
-3. **If NOT installed** → check `.agent-settings.yml` for `personal.rtk_last_asked`. If not today → prompt user (see Installation below).
+1. Read `personal.rtk_installed` from `.agent-settings.yml`.
+2. **If `true`** → prefix commands with `rtk` when output >30 lines expected.
+3. **If `false` or missing** → use plain commands. Do not prompt the user.
 4. After wrapping: verify output is useful (not truncated on completeness-critical commands).
 
+Installation and one-time setup are owned by
+[`/onboard`](../../commands/onboard.md). If the user asks to install rtk
+outside onboarding, follow the Installation section below and set
+`personal.rtk_installed: true` on success.
+
 ## Output format
 
 1. Wrapped command with `rtk` prefix
 2. Token savings estimate (if first use in conversation)
 
-## Detection & Installation
-
-**Before using CLI commands that produce verbose output**, check if rtk is available:
-
-1. Run `which rtk` (silent, no output to user).
-2. **If installed** → prefix commands with `rtk` when output may exceed ~30 lines.
-3. **If NOT installed** → check `.agent-settings.yml` for `personal.rtk_last_asked`:
-   - If missing or date is **before today** → prompt the user (see below).
-   - If date is **today** → skip, don't ask again. Use normal commands.
-
-**Prompt template** (translate to user's language):
-
-> 💡 **rtk** (Rust Token Killer) is not installed on your system.
-> It reduces token consumption by 60-90% on common CLI commands.
->
-> 1. Install via Homebrew — `brew install rtk` (recommended on macOS)
-> 2. Install via Cargo — `cargo install rtk`
-> 3. Skip for now — I'll ask again tomorrow
-
-**On user response:**
-- **1 or 2** → Run the chosen install command. After success:
-  1. `rtk --version` to verify installation.
-  2. `rtk init --global` to enable auto-rewrite hooks.
-  3. Apply **Post-Install Setup** (see below) — telemetry, tee, audit logging.
-  4. Generate project-local filters (see Post-Install Setup).
-  5. Save `personal.rtk_installed: true` in `.agent-settings.yml`.
-- **3** → Save `personal.rtk_last_asked: YYYY-MM-DD` (today) in `.agent-settings.yml`. Use normal commands.
+## Installation (on-demand)
+
+Invoked by `/onboard` when rtk is not on `PATH`, or on explicit user
+request. Never fire unsolicited.
+
+1. Ask which installer to use (macOS → Homebrew; otherwise Cargo).
+2. Run the installer. On success:
+   1. `rtk --version` to verify.
+   2. `rtk init --global` to enable auto-rewrite hooks.
+   3. Apply **Post-Install Setup** below (telemetry, tee, audit logging).
+   4. Generate project-local filters (see Post-Install Setup).
+   5. Write `personal.rtk_installed: true` to `.agent-settings.yml`.
 
 ## Post-Install Setup (mandatory)
 

package/.agent-src/skills/rule-writing/SKILL.md

@@ -98,9 +98,9 @@ the PR or split by responsibility.
 
 * Run `python3 scripts/skill_linter.py .agent-src.uncompressed/rules/{name}.md`
   → must report **0 FAIL**.
-* Run `task sync` to regenerate `.agent-src/rules/{name}.md`.
-* Run `task generate-tools` to project into `.claude/`, `.cursor/`, `.clinerules/`, `.windsurfrules`.
-* Run `task ci` — must exit 0 except for tolerated warnings.
+* Run `bash scripts/compress.sh --sync` to regenerate `.agent-src/rules/{name}.md`.
+* Run `python3 scripts/compress.py --generate-tools` to project into `.claude/`, `.cursor/`, `.clinerules/`, `.windsurfrules`.
+* Run the full CI pipeline locally — each script directly — must exit 0 except for tolerated warnings.
 
 ## Frontmatter shape
 
@@ -118,7 +118,7 @@ source: package           # or project for consumer-local rules
 1. Complete rule file at `.agent-src.uncompressed/rules/{name}.md`
 2. Frontmatter fully populated, no placeholders left
 3. Linter output showing 0 FAIL
-4. Confirmation that `task sync` + `task generate-tools` ran clean
+4. Confirmation that `bash scripts/compress.sh --sync` + `python3 scripts/compress.py --generate-tools` ran clean
 
 ## Gotchas
 
@@ -126,7 +126,7 @@ source: package           # or project for consumer-local rules
 * Defaulting to `always` "just in case" — token cost is real, `auto` is default.
 * Description like "Rule about X" — it must describe *when*, not *what*.
 * Pasting a workflow into a rule — if it has numbered steps, split into a skill.
-* Forgetting to run `task generate-tools` — downstream tools stay stale.
+* Forgetting to run `python3 scripts/compress.py --generate-tools` — downstream tools stay stale.
 * Editing `.agent-src/rules/` or `.augment/rules/` directly — those are generated.
 
 ## Do NOT

package/.agent-src/skills/terragrunt/SKILL.md

@@ -171,14 +171,6 @@ The project uses **devbox** for tool management:
 }
 ```
 
-### Taskfile commands
-
-```bash
-task versions          # Show awscli, Terraform, Terragrunt versions
-task check:awscli      # Verify AWS CLI is configured
-task cache:clear       # Clear .terraform and .terragrunt-cache dirs
-```
-
 ### Quick commands (devbox scripts)
 
 ```bash

package/.agent-src/skills/upstream-contribute/SKILL.md

@@ -165,10 +165,11 @@ The files must:
 ### 6. Run quality gates (if working in the package repo)
 
 ```bash
-task lint-skills                    # 0 FAIL required
-task check-compression              # No errors for this file
-task generate-tools                 # Regenerate symlinks
-task consistency                    # Everything in sync
+python3 scripts/skill_linter.py --all          # 0 FAIL required
+python3 scripts/check_compression.py            # No errors for this file
+python3 scripts/compress.py --generate-tools    # Regenerate symlinks
+bash scripts/compress.sh --check                # .agent-src/ in sync with source
+bash scripts/compress.sh --check-hashes         # All hashes match
 ```
 
 If not in the package repo, note that these checks will run in CI after the PR is created.