refactor: follow the cl-prolog-kit / cl-dataflow-kit rename

docs/notes/cl-weave-advanced-usage-audit.md

@@ -28,12 +28,12 @@ and a dedicated `test-completion-properties.lisp` with 12), four separate
 input-state property files (core, navigation, kill-yank, search, completion),
 autosuggest, prompt rendering, and CPS.
 
-## cl-prolog-query integration — a second suite dedicated to it
+## cl-prolog-kit-query integration — a second suite dedicated to it
 
 `nshell/weave` (`nshell.asd`, `t/weave/`) is a whole second ASDF system,
 separate from `nshell/test`, whose own description is "property-based,
-fixture, benchmark, and cl-prolog-query coverage of the completion engine".
-It depends on `cl-prolog/weave` specifically for this. The weave tests use
+fixture, benchmark, and cl-prolog-kit-query coverage of the completion engine".
+It depends on `cl-prolog-kit/weave` specifically for this. The weave tests use
 `prove`/`assert-fact!`/prolog querying against the completion rulebase
 exported by `nshell.domain.completion` (`#:completes #:describes #:has-flag
 #:command-is ...` -- see `nerima-lisp-package-audit.md`), letting tests write
@@ -60,6 +60,6 @@ test.
 "cl-weave を利用して高度な使い方をしてほしい" is satisfied by breadth, not
 one showcase: mutation testing (the framework's own stated "most advanced
 facility"), property-based trials with custom generators and shrinkers,
-a second whole ASDF system dedicated to cl-prolog-query integration, and
+a second whole ASDF system dedicated to cl-prolog-kit-query integration, and
 table-driven consolidation, all exercised by the existing suite rather than
 being demonstrated once and left unused elsewhere.

docs/notes/coverage-analysis.md

@@ -30,7 +30,7 @@ minimum and target fields have been inspected. A passing minimum with
 - Unit and integration tests cover parsing, expansion, completion, environment,
   process, timeout, redirection, command dispatch, and job management.
 - The weave suite exercises the completion knowledge base through property
-  cases, fixtures, benchmarks, direct Prolog queries, and the `cl-prolog/weave`
+  cases, fixtures, benchmarks, direct Prolog queries, and the `cl-prolog-kit/weave`
   bridge.
 - PTY, terminal, and external-binary paths are verified by the PTY-capable
   runner. The Nix sandbox may skip those environment-dependent cases, so the

docs/notes/nerima-lisp-package-audit.md

@@ -14,9 +14,9 @@ appear unqualified):
 
 | package | role in nshell | evidence |
 |---|---|---|
-| `cl-prolog` | completion knowledge base — facts/rules, `map-prolog-solutions` | `domain/completion/rule-data.lisp` (8 refs) |
+| `cl-prolog-kit` | completion knowledge base — facts/rules, `map-prolog-solutions` | `domain/completion/rule-data.lisp` (8 refs) |
 | `cl-parser-kit` | `$((…))` arithmetic tokenizer + Pratt parser | `:import-from` in `package.lisp`; `domain/expansion/arithmetic.lisp` |
-| `cl-dataflow` | reactive dataflow wiring | 12 refs |
+| `cl-dataflow-kit` | reactive dataflow wiring | 12 refs |
 | `cl-host-kit` | host environment, pathname, and process boundaries | `presentation/repl-environment.lisp`, `infrastructure/terminal/ansi.lisp`, `application/builtin-runtime.lisp` |
 | `cl-boundary-kit` | clock/sleeper boundaries (also under cl-process-kit) | 14 refs |
 | `cl-cli` | argument-vector parsing for `main` | 13 refs |
@@ -32,7 +32,7 @@ The test systems are kept separate from the runtime dependency audit:
 | package | role in nshell's test systems | evidence |
 |---|---|---|
 | `cl-weave` | the test framework for the weave suite | `nshell/weave` |
-| `cl-prolog/weave` | cl-prolog-query coverage of the completion engine | `nshell/weave` |
+| `cl-prolog-kit/weave` | cl-prolog-kit-query coverage of the completion engine | `nshell/weave` |
 
 ## Transitive, not adopted directly
 
@@ -62,6 +62,6 @@ nshell never handles (`cl-json-kit`), or a build-time tool (`paredit-cli`).
 Re-run the usage half of this audit with:
 
 ```
-rg -o '\b(cl-prolog|cl-dataflow|cl-boundary-kit|cl-cli|cl-tty-kit|process-kit|history-kit|cl-concurrent-kit)::?[a-z]' src/ \
+rg -o '\b(cl-prolog-kit|cl-dataflow-kit|cl-boundary-kit|cl-cli|cl-tty-kit|process-kit|history-kit|cl-concurrent-kit)::?[a-z]' src/ \
   | perl -pe 's/:.*$//' | sort | uniq -c | sort -rn
 ```

docs/src/guide/concepts.md

@@ -19,7 +19,7 @@ which keeps the interactive core deterministic and unit-testable.
 ## Completion is a knowledge base, not a table
 
 Completion candidates come from a logic knowledge base compiled into a
-[cl-prolog](https://github.com/nerima-lisp/cl-prolog) rulebase, queried through
+[cl-prolog-kit](https://github.com/nerima-lisp/cl-prolog-kit) rulebase, queried through
 predicates such as `completes`, `describes`, `has-flag`, `command-is`,
 `suggests-dir`, and `suggests-file`.
 

docs/src/guide/recipes.md

@@ -12,7 +12,7 @@ pipeline-graph --mermaid 'cat access.log | grep 404 | wc -l'
 ```
 
 It is a diagnostic built on the
-[cl-dataflow](https://github.com/nerima-lisp/cl-dataflow) computation-graph
+[cl-dataflow-kit](https://github.com/nerima-lisp/cl-dataflow-kit) computation-graph
 toolkit, and it validates the pipeline as it builds the graph, so a malformed
 redirect shows up as an error rather than a picture.
 

docs/src/project/contributing.md

@@ -45,7 +45,7 @@ Iterate with the smallest relevant suite, then run the full gate before review.
 |---|---|
 | Anything | `nix flake check --print-build-logs` before review; the full gate runs on `x86_64-linux` CI |
 | Domain logic, builtins, parser, expansion | `nix run .#test` |
-| Completion engine or cl-prolog knowledge base | `sbcl --script scripts/weave.lisp` |
+| Completion engine or cl-prolog-kit knowledge base | `sbcl --script scripts/weave.lisp` |
 | PTY, subprocess, terminal, signal, job control | the non-sandboxed run below |
 | Coverage-oriented validation | `nix develop -c sbcl --script scripts/coverage.lisp` |
 

docs/src/project/public-readiness.md

@@ -50,7 +50,7 @@ What is actually checkable on a plain `aarch64-darwin` development machine
   `nix build --no-link .#checks.aarch64-darwin.formatting` both exit 0 -- the
   mkdocs `--strict` build and the treefmt gate pass on this platform.
 - The Darwin `build`, `default`, and `smoke-test` check attributes can remain
-  unavailable when the pinned `cl-prolog` package has no Darwin build. The
+  unavailable when the pinned `cl-prolog-kit` package has no Darwin build. The
   complete `nix flake check` and release-bundle gate are therefore verified on
   `x86_64-linux` CI; run available Darwin checks directly rather than treating a
   missing attribute as a feature failure.

docs/src/reference/architecture.md

@@ -59,7 +59,7 @@ layer where it fits the domain-driven design:
   rule-based tokenizer and Pratt (operator-precedence) parser drive `$((...))`
   arithmetic, which parses to an AST and then evaluates, adding `**`, bitwise
   `& | ^ ~`, shifts `<< >>`, and the ternary `?:`.
-- **[cl-dataflow](https://github.com/nerima-lisp/cl-dataflow)** — renders
+- **[cl-dataflow-kit](https://github.com/nerima-lisp/cl-dataflow-kit)** — renders
   pipelines as validated computation graphs (`pipeline-graph`) and models the
   job lifecycle as an analyzable state machine.
 - **[cl-boundary-kit](https://github.com/nerima-lisp/cl-boundary-kit)** — makes
@@ -83,7 +83,7 @@ layer where it fits the domain-driven design:
   timeout-guarded process launch, escalating SIGTERM to SIGKILL across a
   child's whole process group so a timed-out command substitution leaves no
   orphaned descendants.
-- **[cl-prolog](https://github.com/nerima-lisp/cl-prolog)** — the logic engine
+- **[cl-prolog-kit](https://github.com/nerima-lisp/cl-prolog-kit)** — the logic engine
   behind the completion knowledge base.
 
 ## Test suites
@@ -93,9 +93,9 @@ exposed as Nix checks:
 
 - **`nshell/test`** — the primary regression suite, in `t/`.
 - **`nshell/weave`** — a focused suite exercising the completion engine's
-  cl-prolog knowledge base with property-based tests, fixtures, benchmarks, and
+  cl-prolog-kit knowledge base with property-based tests, fixtures, benchmarks, and
   direct Prolog queries (`findall`, negation-as-failure, foreign predicates)
-  plus the `cl-prolog/weave` query bridge.
+  plus the `cl-prolog-kit/weave` query bridge.
 
 Cases that need a real PTY, `stty`, or external binaries cannot run in the Nix
 sandbox and are covered by CI's separate `integration` job; see