@event4u/agent-config 5.6.1 → 5.8.0

package/.agent-src/commands/prediction-pool.md

@@ -0,0 +1,234 @@
+---
+model_tier: inherit
+name: prediction-pool
+tier: 2
+description: "Fill a prediction pool (kicktipp, football/basketball WM): optimize expected points under the rules, enter tips via Playwright. Triggers 'Tippspiel', 'kicktipp', 'predict the pool'."
+skills: [prediction-pool-optimizer]
+suggestion:
+  eligible: true
+  trigger_description: "fill my Tippspiel, kicktipp tips, predict the pool, betting/prediction pool predictions, optimize my pool tips for an event"
+  trigger_context: "user wants tips for a prediction pool (kicktipp etc.) for an upcoming sports event and wants them computed and/or entered into the pool's web UI"
+workspaces:
+  - small-business
+packs:
+  - fun
+lifecycle: experimental
+trust:
+  level: experimental
+install:
+  default: false
+  removable: true
+---
+
+# /prediction-pool
+
+`/prediction-pool [<event>] [--pool-url <url>] [--council off|event|match] [--submit] [--fresh|--continue]`
+
+Fills out a **prediction pool** (kicktipp-style company pools: football
+WM, basketball WM, …) by optimizing for the **maximum expected points
+under the pool's actual scoring rules** — not for the most likely match
+outcome. Researches market odds, computes expected-value tips, presents a
+table for approval, then enters them into the pool's web UI via Playwright.
+
+The cognition lives in the [`prediction-pool-optimizer`](../skills/prediction-pool-optimizer/SKILL.md)
+skill (rules → market odds → expected value → participant field → tip).
+This command is the orchestrator: event selection, the persisted analysis,
+the optional AI-council pass, the Playwright entry, and the human gates.
+
+**Hard gates (always):**
+- **You log in.** The agent opens the page headful; it never reads,
+  stores, or types your credentials.
+- **The agent never submits.** It fills the candidate tips and stops —
+  *you* submit, unless you passed `--submit` (or say so this turn).
+- **Not betting or financial advice.** Decision support for a game; you
+  approve and submit.
+
+**Block on ambiguity:** an unresolvable event, a missing pool URL with no
+saved analysis, or contradictory flags (`--fresh` and `--continue`,
+`--council` value not in `off|event|match`) halts with a precise message —
+no silent best-guess.
+
+## Inputs
+
+| Input | Required | Meaning |
+|---|---|---|
+| `<event>` | no | The event to tip (e.g. "Football WM 2026"). Omitted → the command proposes 1–3 upcoming events + free input (Step 1). |
+| `--pool-url <url>` | once per event | The pool's tip page. Saved into the per-event analysis; reused on later runs. |
+| `--council off\|event\|match` | no | AI-council depth for the analysis. Default `off`; the command offers it at Step 4. `event` = one pass over the whole sheet; `match` = per-match (costlier). |
+| `--submit` | no | Pre-authorize the agent to submit after entry. Default: never submit. |
+| `--fresh` / `--continue` | no | Force a new analysis vs. build on the saved one (Step 2). Default: ask. |
+
+## Steps
+
+### 1. Resolve the event (block on ambiguity)
+
+- `<event>` supplied → use it; derive a slug (`football-wm-2026`).
+- **Omitted → propose, then stop and wait.** Name 1–3 *upcoming* events
+  (use the current date; research what is imminent) and offer free input:
+
+  ```
+  > Which event should I generate tips for?
+  >
+  > 1. <imminent event A>
+  > 2. <imminent event B>
+  > 3. <imminent event C>
+  > 4. Something else — name it
+  ```
+
+### 2. Resolve the per-event analysis (state ground-truth)
+
+Look for `agents/tmp/prediction-pool/<slug>.md`.
+
+- **None →** start a new analysis (created at Step 9, never before).
+- **Exists →** read it. Then:
+  - It was **edited externally** since the last run (mtime newer than the
+    last `## Run <ts>` header, or it no longer parses) → surface that and
+    ask **merge / overwrite / abort** — never silently overwrite manual edits.
+  - Otherwise ask **build on it** (`--continue`) or **start fresh**
+    (`--fresh`); the flags pre-answer this for non-interactive runs.
+
+The analysis holds the pool URL, the parsed rules, the current tip state,
+and standing notes — so re-tuning later is fast.
+
+### 3. Read the pool rules FIRST
+
+Resolve the pool URL (`--pool-url` → saved analysis → ask). Then run the
+optimizer skill's **rules pass** against the pool's rule page: exact result
+vs. goal-difference vs. tendency points, bonus questions, joker rules,
+quote/rarity scoring, special rules, deadlines, strategy limits. **Enumerate
+every open question as an explicit checklist** (scores AND every bonus /
+award / special question) — that list is the run's contract; each entry must
+reach an answer. **No tips before the rules are understood** — the whole
+strategy depends on them.
+
+### 4. Run the optimizer + offer the AI-council pass
+
+Run [`prediction-pool-optimizer`](../skills/prediction-pool-optimizer/SKILL.md),
+adapted to the event's sport (football / basketball / …): a **de-vigged
+consensus across the 5–10 biggest publicly-viewable bookmakers**
+(sharp-weighted, never one portal) as the primary signal → expected value
+under the Step-3 rules → participant field → tip. **Answer every entry on the
+Step-3 checklist** — scores AND every bonus / award / special question (top
+scorer, "team of the top scorer", group winners, most cards …); none blank.
+Tournament/outright/award probabilities come from real markets **or** the
+skill's executed Poisson helper — **never** a hallucinated "I simulated
+10,000 runs".
+
+**Offer the AI-council pass (default off).** Unless `--council` is set,
+ask once:
+
+```
+> Run the AI council over the analysis for a sharper second opinion?
+>
+> 1. No — use my analysis as is
+> 2. Yes, per event — one council pass over the whole tip sheet
+> 3. Yes, per match — each match judged separately (more accurate, costlier)
+```
+
+When on, run **graduated** to control cost: a cheap single-model pass
+flags the riskiest matches first, then the full council reviews only the
+flagged matches (`event`) or every match (`match`). Fold the council's
+verdicts back into the table before Step 5. Council spend always asks
+first per [`ai-council`](../../../core/.agent-src.uncondensed/skills/ai-council/SKILL.md).
+
+### 5. Output the approval table — ask whether & where to enter
+
+Present the tips exactly as they would be entered, then **wait**:
+
+| Match | Tip | Prob / EV | Risk | 1-line reason | Books used |
+|---|---|---|---|---|---|
+
+Follow with group standings, the full bracket, **and a bonus/special-answer
+table with one row per open question from the Step-3 checklist — every entry
+answered, none blank**:
+
+| Question | Answer | Prob / EV | Risk | 1-line reason | Source (market / model) |
+|---|---|---|---|---|---|
+
+Then ask **whether to enter** and **into which pool** (the saved URL, a
+different one, or none). Do **not** write the analysis yet (Step 9) — tips
+are not yet confirmed.
+
+### 6. Enter via Playwright (you log in)
+
+Open the pool page **headful**. **You log in yourself** — the agent waits
+and never touches credentials. Resolve the platform adapter:
+
+- **Known platform** (e.g. kicktipp) → load the declarative selector map
+  `scripts/prediction-pool/adapters/<platform>.yml` (field → CSS selector). The
+  generic, trusted driver fills the inputs from the map — the adapter is
+  **data, not code**.
+- **Unknown platform → vision-assisted synthesis.** Screenshot the page,
+  identify the tip inputs, **highlight them** and ask you to confirm the
+  mapping, then fill from the confirmed ephemeral map. No code is run from
+  an untrusted source.
+
+### 7. Stop before submit
+
+Fill the candidate results and **stop**. *You* press submit. The agent
+submits only if `--submit` was passed or you authorize it **this turn**
+(mirrors [`non-destructive-by-default`](../../../core/.agent-src.uncondensed/rules/non-destructive-by-default.md)).
+
+### 8. Offer a second pool
+
+Ask whether to also enter the same (or re-optimized) tips on another
+pool / site. If yes, loop Steps 3–7 for that pool.
+
+### 9. Persist / extend the analysis (only now)
+
+Append a run-stamped section to `agents/tmp/prediction-pool/<slug>.md`: pool
+URL(s), parsed rules, the entered tips with state `entered, not submitted`
+(or `submitted` if Step 7 submitted), council verdicts if any, and
+standing notes. Append-only — earlier runs stay as history. This is the
+base the next run reads (Step 2).
+
+### 10. New-platform adapter — offer to contribute (gated)
+
+If Step 6 synthesized a new selector map, offer to (a) save it locally and
+(b) open a **PR** adding `scripts/prediction-pool/adapters/<platform>.yml` so
+coverage grows for everyone. The PR carries **declarative selector data
+only** — never executable code — and only on explicit permission (no
+auto-commit, no auto-push).
+
+### 11. Report
+
+Print: event + slug, pool URL(s), rules summary, council depth used,
+matches tipped, entry result (`entered, not submitted` | `submitted`),
+adapter (`<platform>.yml` | `vision-synthesized` | `pr-offered`), analysis
+file path. No commit. No push.
+
+## Rules
+
+- **You log in; the agent never handles credentials.** Headful only.
+- **The agent never submits** unless `--submit` or this-turn authorization.
+- **Rules first.** No tips before the pool's scoring is parsed.
+- **Answer every open question.** Scores AND every bonus / award / special
+  question; a run that ships scorelines only and leaves bonus questions blank
+  is incomplete.
+- **Consensus odds, not one portal.** Build the base from a de-vigged,
+  sharp-weighted consensus across the 5–10 biggest viewable books — never
+  mirror a single bookmaker.
+- **No hallucinated simulation.** Outright odds or executed Poisson code —
+  never a claimed-but-unrun Monte-Carlo.
+- **Analysis is written only after tips are confirmed** (Step 9), never
+  before — and never silently over an externally edited file.
+- **Adapters are declarative data, not code.** Unknown platforms use the
+  ephemeral vision path; contributed adapters are selector maps via PR.
+- **AI council is opt-in, default off**, and always asks before spending.
+- **Not betting or financial advice.** A fun tool; you decide and submit.
+- **No commit, no push, no PR** without explicit permission (the adapter
+  PR offer is gated).
+- **Kill-switch.** Ships `lifecycle: experimental` · `install.default:
+  false`. Disable = remove the command + `prediction-pool-optimizer` skill, then
+  regenerate the projected tool trees.
+
+## See also
+
+- [`prediction-pool-optimizer`](../skills/prediction-pool-optimizer/SKILL.md) — the
+  rules → odds → EV → field → tip cognition.
+- [`scripts/prediction-pool/adapters/_schema.md`](../../../scripts/prediction-pool/adapters/_schema.md) —
+  the declarative adapter data contract (for PR contributions).
+- [`ai-council`](../../../core/.agent-src.uncondensed/skills/ai-council/SKILL.md) —
+  the optional second-opinion pass (Step 4).
+- [`playwright-architect`](../../../core/.agent-src.uncondensed/skills/playwright-architect/SKILL.md) —
+  browser-automation patterns for the entry step.

package/.agent-src/commands/profile/activate.md

@@ -0,0 +1,81 @@
+---
+model_tier: medium
+name: profile:activate
+tier: 2
+cluster: profile
+sub: activate
+description: Activate a session profile — surface only the named profile/pack closure plus core artefacts, no persistence
+suggestion:
+  eligible: true
+  trigger_description: "activate the laravel/po/finance profile for this session, switch my surface to X, only show me X commands"
+  trigger_context: "user wants the surfaced commands/skills narrowed to one audience for the current session"
+workspaces:
+  - agent-config-maintainer
+packs:
+  - meta
+---
+
+# /profile activate
+
+Activate one or more session profiles / packs. Writes the expanded
+`runtime.active_packs` overlay and re-biases the surfaced set for the rest
+of the session.
+
+## Instructions
+
+### 1. Resolve + write the overlay
+
+Run the helper (it resolves aliases ↔ pack ids, fails fast on a
+not-installed pack, expands the `requires_hint` closure, and writes the
+overlay **atomically**):
+
+```bash
+python3 -m scripts.config.session_profiles activate <name…> --json
+```
+
+- **Exit 0** → parse the JSON: `active_packs`, `closure_added`, `notes`.
+- **Exit 2** (`error:` on stderr) → the token is unknown or the pack is
+  not installed. Surface the error verbatim and stop — do **not** write a
+  partial overlay. Suggest `/profile show` or naming an installed pack.
+
+### 2. Confirm the switch
+
+Announce, using the helper's output:
+
+```
+> profile active: {active_packs}
+> surfaced: {N} commands, {M} skills ({H} hidden behind inactive packs)
+```
+
+Get the counts from `python3 -m scripts.config.session_profiles show --json`.
+
+### 3. Re-bias the surface for the rest of the session
+
+From now until `/profile deactivate` (or a new session), when the user asks
+"what can I do" / you list commands or skills, surface **only** the active
+set. Compute it with:
+
+```bash
+python3 -m scripts.config.session_profiles surface --json
+```
+
+`shown` = surface these; `hidden` = do not list them proactively. Core /
+unscoped artefacts are always in `shown`.
+
+**Execution is NOT gated.** If the user invokes a command/skill that is in
+`hidden`, run it anyway and prepend one line:
+
+```
+> note: `{name}` is from an inactive pack ({packs}); `/profile activate {pack}` to surface it.
+```
+
+## Gotchas
+
+- Never write to `.agent-settings.yml` — the overlay is local-only.
+- Multiple names union: `/profile activate laravel po`.
+- A corrupt overlay is ignored (fail-open) → re-run activate to repair it.
+
+## See also
+
+- [`/profile show`](show.md) · [`/profile deactivate`](deactivate.md)
+- [`docs/contracts/session-profile-overlay.md`](../../../docs/contracts/session-profile-overlay.md)

package/.agent-src/commands/profile/deactivate.md

@@ -0,0 +1,68 @@
+---
+model_tier: medium
+name: profile:deactivate
+tier: 2
+cluster: profile
+sub: deactivate
+description: Deactivate the session profile — clear the overlay (or drop named packs) so the full surface returns
+suggestion:
+  eligible: true
+  trigger_description: "deactivate the profile, clear my session profile, show everything again, reset the surface"
+  trigger_context: "user wants the full command/skill surface back after a profile was active"
+workspaces:
+  - agent-config-maintainer
+packs:
+  - meta
+---
+
+# /profile deactivate
+
+Clear the session-profile overlay (or remove only named packs). This is the
+**explicit** deactivation trigger locked in Phase 0 (option a) — there is
+no silent session reset.
+
+## Instructions
+
+### 1. Clear or narrow
+
+- `/profile deactivate` (no args) → clear the whole overlay:
+
+  ```bash
+  python3 -m scripts.config.session_profiles deactivate --json
+  ```
+
+- `/profile deactivate <name…>` → remove only the **named packs** from the
+  active set (never their closure), so a shared dependency survives:
+
+  ```bash
+  python3 -m scripts.config.session_profiles deactivate <name…> --json
+  ```
+
+Parse `active_packs` from the JSON.
+
+### 2. Confirm
+
+```
+> profile cleared — full surface restored.
+```
+
+or, when packs remain:
+
+```
+> active now: {active_packs} — full surface for everything else.
+```
+
+### 3. Restore the full surface
+
+Resume surfacing every command/skill from this point in the session.
+
+## Gotchas
+
+- Deactivation is idempotent — clearing an empty overlay is a no-op.
+- Only named packs are removed: `/profile deactivate laravel` leaves `php`
+  and `engineering-base` active (they are their own entries in the overlay).
+  Removing a pack only ever widens the surface, never hides something.
+
+## See also
+
+- [`/profile activate`](activate.md) · [`/profile show`](show.md)

package/.agent-src/commands/profile/show.md

@@ -0,0 +1,70 @@
+---
+model_tier: medium
+name: profile:show
+tier: 2
+cluster: profile
+sub: show
+description: Show the active session profile — active packs and surfaced/hidden command+skill counts (observability surface)
+suggestion:
+  eligible: true
+  trigger_description: "show my active profile, which packs are active, what's surfaced this session, profile status"
+  trigger_context: "user wants to see the current session-profile state and what it hides"
+workspaces:
+  - agent-config-maintainer
+packs:
+  - meta
+---
+
+# /profile show
+
+Observability surface for the session-profile overlay — the diagnostic the
+council required so a "wrong commands in /help" symptom is one command away
+from its cause.
+
+## Instructions
+
+### 1. Read the state
+
+```bash
+python3 -m scripts.config.session_profiles show --json
+```
+
+Fields: `active_packs`, `commands_shown`, `skills_shown`, `hidden_total`.
+
+### 2. Report
+
+- **No overlay** (`active_packs` empty):
+
+  ```
+  > no profile active — full surface (everything shown).
+  ```
+
+- **Overlay active:**
+
+  ```
+  > active packs: {active_packs}
+  > surfaced: {commands_shown} commands, {skills_shown} skills
+  > hidden behind inactive packs: {hidden_total}
+  > clear with /profile deactivate
+  ```
+
+### 3. (Optional) list the split
+
+If the user asks *which* artefacts are hidden, run:
+
+```bash
+python3 -m scripts.config.session_profiles surface --json
+```
+
+and list `hidden[]` (name + packs). Keep it scannable; group by pack if long.
+
+## Gotchas
+
+- `show` never writes — pure read, safe to run anytime.
+- A corrupt overlay reads as empty (fail-open); if `show` reports "no
+  profile" right after an `activate`, the overlay file is malformed —
+  re-run `/profile activate`.
+
+## See also
+
+- [`/profile activate`](activate.md) · [`/profile deactivate`](deactivate.md)

package/.agent-src/commands/profile.md

@@ -0,0 +1,68 @@
+---
+model_tier: medium
+name: profile
+tier: 2
+description: Session-profile orchestrator — activate / deactivate / show the active packs for this session (recommendation-bias surface filter, no persistence)
+cluster: profile
+type: orchestrator
+suggestion:
+  eligible: true
+  trigger_description: "activate a profile for this session, switch to the laravel/po/finance surface, show which packs are active, deactivate the profile"
+  trigger_context: "user wants to narrow the surfaced commands/skills to one audience for the current session without changing committed config"
+workspaces:
+  - agent-config-maintainer
+packs:
+  - meta
+---
+
+# /profile
+
+Top-level orchestrator for the `/profile` family. Activates a **session
+profile** — an ephemeral `runtime.active_packs` overlay that biases which
+commands/skills are *surfaced* this session, without persisting anything
+to committed config.
+
+The overlay is a runtime modulation of the existing `pack` axis, **not** a
+fifth axis (ADR-010 addendum, 2026-06-02). It is written to
+`.agent-settings.local.yml` (in `agents/settings/`, gitignored, deepest
+layer), never the committed `.agent-settings.yml`.
+
+## Sub-commands
+
+| Sub-command | Routes to | Purpose |
+|---|---|---|
+| `/profile activate <name…>` | `commands/profile/activate.md` | Surface only the named profile/pack closure + core artefacts |
+| `/profile deactivate [name…]` | `commands/profile/deactivate.md` | Clear the overlay (or drop named packs) → full surface returns |
+| `/profile show` | `commands/profile/show.md` | Active packs + surfaced/hidden counts (observability) |
+
+`<name>` is a session-profile alias (`developer`, `po`, `finance`, `gtm`,
+`content`) from `config/discovery/session-profiles.yml` **or** a raw pack
+id from `config/discovery/packs.yml` (`laravel`, `php`, …). Multiple names
+union their closures: `/profile activate laravel po`.
+
+## Key facts (locked in Phase 0)
+
+- **Recommendation-bias only.** Activation filters the *surfaced* set
+  (`/help` + `<available_skills>`); execution is **never gated** — an
+  inactive-pack command still runs, with a one-line notice.
+- **Explicit deactivation.** The overlay survives an IDE restart; clear it
+  with `/profile deactivate`. A new session emits a staleness *notice*
+  (never a silent reset — the registry-refresh Catch-22, see
+  `agents/settings/contexts/session-host-capability-audit.md`).
+- **Fail-open.** A corrupt overlay is ignored → the full surface returns.
+- **Kill-switch.** Delete `runtime.active_packs` from the local file (or
+  the file itself) to reset.
+
+## Dispatch
+
+1. Parse `/profile <sub-command> [args]`.
+2. Look up the sub-command above; load that file and follow its
+   `## Instructions` verbatim with the remaining args.
+3. Bare `/profile` (no sub-command) → run `/profile show`.
+4. Unknown sub-command → print the table and ask which the user meant.
+
+## See also
+
+- [`mode`](mode.md) — the persist-to-disk precedent this command deliberately avoids.
+- [`docs/decisions/ADR-010-profile-pack-preset-boundary.md`](../../docs/decisions/ADR-010-profile-pack-preset-boundary.md) — axis boundary + overlay addendum.
+- [`docs/contracts/session-profile-overlay.md`](../docs/contracts/session-profile-overlay.md) — overlay schema + surface-filter contract.

package/.agent-src/commands/project-health.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: project-health
 tier: 1
 skills: [quality-tools]

package/.agent-src/commands/quality-fix.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: quality-fix
 tier: 1
 skills: [quality-tools]

package/.agent-src/commands/roadmap/process-full.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: roadmap:process-full
 tier: 2
 cluster: roadmap

package/.agent-src/commands/roadmap/process-phase.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: roadmap:process-phase
 tier: 2
 cluster: roadmap

package/.agent-src/commands/roadmap/process-step.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: roadmap:process-step
 tier: 2
 cluster: roadmap

package/.agent-src/commands/roadmap.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: roadmap
 tier: 1
 description: Roadmap orchestrator — routes to create (authoring) and process-step / process-phase / process-full (autonomous execution).

package/.agent-src/commands/set-cost-profile.md

@@ -1,8 +1,8 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: set-cost-profile
 tier: 1
-description: Change the cost_profile in .agent-settings.yml — shows each profile's meaning and applies the selection
+description: Change the rule_loading_tier in .agent-settings.yml — shows each profile's meaning and applies the selection
 skills: [file-editor]
 suggestion:
   eligible: false
@@ -17,7 +17,7 @@ packs:
 
 # /set-cost-profile
 
-Changes `cost_profile` in `.agent-settings.yml`. Four profiles are defined in
+Changes `rule_loading_tier` in `.agent-settings.yml`. Four profiles are defined in
 the [`agent-settings` template](../templates/agent-settings.md#cost-profiles):
 
 - `minimal` · `balanced` · `full` · `custom`
@@ -32,7 +32,7 @@ the [`agent-settings` template](../templates/agent-settings.md#cost-profiles):
   directly or ask the agent — the merge rules live in
   [`layered-settings`](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules).
 - For role modes use [`/mode`](mode.md) — different concept (sets
-  `roles.active_role`, not `cost_profile`).
+  `roles.active_role`, not `rule_loading_tier`).
 
 ## Steps
 
@@ -49,7 +49,7 @@ Profile names are case-insensitive on input; the file value stays lowercase.
 Read `.agent-settings.yml`. If missing, tell the user to run
 `scripts/install` first and stop — do not create the file here.
 
-Extract the current `cost_profile` value.
+Extract the current `rule_loading_tier` value.
 
 ### 3. Load profile descriptions
 
@@ -63,7 +63,7 @@ Render the current value and present numbered choices with the hint text
 extracted in step 3:
 
 ```
-> Current: cost_profile = {current}
+> Current: rule_loading_tier = {current}
 >
 > 1. minimal — {hint from template}
 > 2. balanced — {hint from template}
@@ -77,7 +77,7 @@ value directly — still echo the old → new line in step 6.
 
 ### 5. Write the value
 
-Update `cost_profile` in `.agent-settings.yml` using the
+Update `rule_loading_tier` in `.agent-settings.yml` using the
 [section-aware merge rules](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules)
 (preserve comments, preserve key order, touch only the changed field).
 
@@ -86,7 +86,7 @@ If the user picked "Keep current", do nothing and stop.
 ### 6. Confirm
 
 ```
-> cost_profile: {old} → {new}
+> rule_loading_tier: {old} → {new}
 ```
 
 If the new profile activates a surface the user hasn't used before
@@ -106,7 +106,7 @@ steps here, that's the docs' job.
 ## Cloud Behavior
 
 On cloud surfaces (Claude.ai Web, Skills API) this command is **fully inert** —
-there is no `.agent-settings.yml` to write and no `cost_profile` toggle to
+there is no `.agent-settings.yml` to write and no `rule_loading_tier` toggle to
 flip. Cost behaviour on those surfaces is governed by the platform itself.
 
 ## See also

package/.agent-src/commands/skill/preview.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: skill:preview
 tier: 2
 cluster: skill
@@ -46,8 +46,8 @@ plain-language summary. `--format json` is machine-readable.
 
 ### 3. Present the summary
 
-Show the plain-language preview: the skill's execution type (a manual-execution skill
-renders **"instructional only — no automatic execution"**; an assisted-execution skill
+Show the plain-language preview: the skill's execution type (a `manual` skill
+renders **"instructional only — no automatic execution"**; an `assisted` skill
 renders its proposed actions), declared steps, tools, and any file/command
 targets. End on the contract reminder that preview shows *declared intent*, not
 a guarantee of side-effect-freeness.

package/.agent-src/commands/skill.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: skill
 tier: 2
 description: Single-skill orchestrator — routes to preview. Non-destructive "what will this skill do?" before you run it.

package/.agent-src/commands/skills/discover.md

@@ -1,5 +1,5 @@
 ---
-model_tier: inherit
+model_tier: medium
 name: skills:discover
 tier: 2
 cluster: skills