docs: retire construction-phase notes

AGENTS.md

@@ -14,21 +14,14 @@ and composing their model. Treat domain adaptation as a first-class task, not a
 **Interaction principle.** When a decision genuinely depends on something you don't know,
 ask one focused question — never default to the longest possible explanation.
 
-> **Build status.** This assistant is under active construction
-> ([autofit_assistant#1](https://github.com/PyAutoLabs/autofit_assistant/issues/1)). Steps
-> below that reference `autoassistant/` tooling, `af_*` skills, or `wiki/core/` pages apply
-> **once those files exist**; until then, note the gap to the user instead of improvising a
-> replacement.
-
 ## Session start — do this first, every session
 
 1. **Maintainer mode.** Check for `.maintainer`; if present, read `modes/maintainer.md`.
    (`touch`/`rm .maintainer`; gitignored.)
 2. **User profile.** Read `wiki/project/profile.md` when present and use it to calibrate depth
    and domain context. Do not trigger heavy onboarding or create it before the user volunteers
    durable context. *(Skipped in maintainer mode.)*
-3. **Environment + API drift-check** *(only in a session that will generate or run code, and
-   only once `autoassistant/` exists)*:
+3. **Environment + API drift-check** *(only in a session that will generate or run code)*:
    ```bash
    python autoassistant/audit_skill_apis.py --check-version
    ```
@@ -230,7 +223,6 @@ Load operational references on demand, not every session:
 
 - **Science projects.** `autofit_assistant` is the copilot; a science project is a separate
   repo created and managed through the `start-new-project` skill.
-- **Installation / environment** → `wiki/core/operations/` (once Phase 2 lands; until then
-  the rule of thumb is `pip install autofit` in a project venv).
+- **Installation / environment** → `wiki/core/operations/`.
 - **External resources** (HowToFit, RTD, `autofit_workspace`) + audience routing →
   [`skills/_style.md`](./skills/_style.md) "Adaptive depth".

README.md

@@ -11,12 +11,6 @@ data and your model, hand it your likelihood code, or discuss inference theory 
 assistant explains the statistics as it goes and writes runnable PyAutoFit Python
 workflows that stay in your repo.
 
-> 🚧 **Under construction.** The assistant is being built in public
-> ([roadmap: issue #1](https://github.com/PyAutoLabs/autofit_assistant/issues/1)).
-> Landed so far: the constitution (`AGENTS.md`), modes, the skills framework, and the wiki
-> scaffolding. Coming next: the workspace-paired `af_*` skills, the core statistics wiki,
-> the domain-adaptation flow, worked demos, and benchmarks.
-
 ## What makes this assistant different
 
 Domain assistants like [`autolens_assistant`](https://github.com/PyAutoLabs/autolens_assistant)

modes/maintainer.md

@@ -28,10 +28,8 @@ skills, wiki schema, hooks, or infrastructure — not user inference work. `AGEN
 Use the existing skills, not new docs:
 
 - Authoring or evolving a skill → [`skills/_bootstrap_skill.md`](../skills/_bootstrap_skill.md).
-- Regenerating `wiki/core/` against pinned sources → `af_update_wiki` *(lands with the
-  core wiki — Phase 2 of [autofit_assistant#1](https://github.com/PyAutoLabs/autofit_assistant/issues/1))*.
-- API gate / version baseline → the `af_audit_skill_apis` skill *(lands with the
-  `autoassistant/` tooling — Phase 1)*.
+- Regenerating `wiki/core/` against pinned sources → `af_update_wiki`.
+- API gate / version baseline → [`skills/af_audit_skill_apis.md`](../skills/af_audit_skill_apis.md).
 
 ## Relationship to autolens_assistant (the reference implementation)
 
@@ -50,9 +48,9 @@ assistant respects. When maintaining this repo:
 
 ## Release-time wiki-currency check
 
-The currency rules (symbol audit, idiom deny-list, provenance) will live in exactly one
-place — a `wiki-currency.yml` workflow in this repo driving
-`autoassistant/audit_skill_apis.py`, invoked both by PyAutoBuild releases
-(`workflow_call`) and on every PR — mirroring the autolens_assistant contract. *(Lands in
-Phases 1/4; until then there is no automated currency gate — check manually against the
-installed stack before shipping wiki or skill edits.)*
+The currency rules (symbol audit, idiom deny-list, provenance, citation paths, version
+drift) live in exactly one place —
+[`.github/workflows/wiki-currency.yml`](../.github/workflows/wiki-currency.yml), driving
+`autoassistant/audit_skill_apis.py` — and run on every PR (plus `workflow_dispatch`, and
+`workflow_call` for PyAutoBuild releases once the release-side wiring lands). When you
+change the rules, edit them there only.

skills/README.md

@@ -25,10 +25,8 @@ configured) via symlinks; the canonical files live here.
 
 ## Index
 
-> 🚧 The `af_*` skill set is being built
-> ([roadmap: issue #1](https://github.com/PyAutoLabs/autofit_assistant/issues/1)).
-> This index lists what exists now; the planned set (domain adaptation, search
-> chaining, simulation, debugging) lands phase by phase.
+> A follow-up tranche (`af_chain_searches`, `af_custom_analysis`, `af_simulate_dataset`,
+> `af_plot_fit`, `af_debug_fit_failure`) is planned; everything below is complete.
 
 ### Meta
 

wiki/core/README.md

@@ -7,13 +7,6 @@ two siblings — [`../literature/`](../literature/) and [`../project/`](../proje
 the user's domain papers and a per-clone journal respectively (see
 [`../README.md`](../README.md) for the overview).
 
-> 🚧 The reference pages land in Phase 2 of
-> [autofit_assistant#1](https://github.com/PyAutoLabs/autofit_assistant/issues/1):
-> priors and model composition, one page per shipped sampler, nested sampling, MCMC/HMC,
-> initialization and chaining, graphical models and expectation propagation, evidence
-> and model comparison, and posterior/aggregator analysis. Until then, derive API truth
-> from the installed source (`AGENTS.md` "Source-of-truth resolution").
-
 ## Organisation
 
 - [`index.md`](./index.md) — top-level map; the entry point for an agent or human

wiki/literature/AGENTS.md

@@ -158,5 +158,4 @@ than guessing.
 This wiki *is* the domain-adaptation surface. When the user's project needs a paper that
 is not here, use the `af_ingest_paper` skill. Add the verified BibTeX metadata, a compact
 source entry, relevant cross-links, and a log row; then run
-`python -m autoassistant.literature validate-citations` *(tooling lands in Phase 1 of
-[autofit_assistant#1](https://github.com/PyAutoLabs/autofit_assistant/issues/1))*.
+`python -m autoassistant.literature validate-citations`.

wiki/literature/bibliography/README.md

@@ -19,5 +19,4 @@ Canonical citation metadata for the literature wiki.
 - Never record local PDF paths. PDFs stay outside the repo (the gitignored `papers/`
   convention), and pages cite by public reference only.
 
-Validation (`python -m autoassistant.literature validate-citations`) lands with the
-Phase 1 tooling ([autofit_assistant#1](https://github.com/PyAutoLabs/autofit_assistant/issues/1)).
+Validate with `python -m autoassistant.literature validate-citations`.