[aitrawog]

Summary

charly candy params plugin-feature --check — the drift gate this repo's candy binds and its CI runs — FAILS on
main: the committed projection is behind its own schema: the schema's doc comments were revised and the
generated types still carry the older text. The pinned generator (charly's provisioned cue
v0.16.1, cue exp gengotypes) carries those comments into the generated file. This PR is the
regeneration: the committed projection is replaced with exactly what the schema now generates.

 candy/plugin-feature/params/cue_types_gen.go | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

No hand edits: the file is the generator's output, byte for byte.

The branch was preserved during a branch/worktree backlog cleanup and is being upstreamed unchanged: its
own commit is 33f5109bbb8a8a4099f2bf40cb4b9a05d103871f — authored before this session, never amended, rewritten or force-pushed. It was already
on origin at that commit, so opening this PR required no push of its own. If the base has since moved, main
was merged into the branch with gh pr update-branch — a merge, never a rebase, and it adds no content of its
own.

How tested

Own gate, PR head — rc=0:

$ charly -C <root>/.worktrees/drift-113/plugin-feature candy params plugin-feature --check
rc=0

Negative control, base — rc=1, which is the defect this PR removes. The base-side command runs against origin/main's blob (e4845ee), because the local
checkout sits on an older pin and the question is about the BASE, not the pin.

$ charly -C <root>/plugin-feature candy params --schema candy/plugin-feature/schema --pkg params --out /tmp/main.go --check
charly: error: command "candy": ... is STALE: it differs from what .../schema generates — run `charly candy params plugin-feature` and commit the result
rc=1

Repo-declared build gate (AGENTS.md), same head — all rc=0:

$ (cd candy/plugin-feature && go build ./...)
rc=0
$ (cd candy/plugin-feature && gofmt -l .)
rc=0
$ (cd candy/plugin-feature && go vet ./...)
rc=0
$ (cd candy/plugin-feature && go test ./...)
ok  	github.com/opencharly/plugin-feature/candy/plugin-feature	0.018s
rc=0

$ charly -C <worktree> box validate
charly box validate: OK — checked 1 candy, 1 deploy, 7 distros, 7 builders; 0 warnings, 0 errors
rc=0

Rulebook compliance

• R0 — skills loaded before the first git/PR action (git-workflow), and this repo's own AGENTS.md read first: its Build / validate / test section is the gate pasted above.
• R1 — no finding to classify: every command above returns 0 with zero warnings; the base-side rc=1 is the defect this PR removes, and it is fixed here rather than filed.
• R2 — no duplication: the projection is generated from candy/plugin-feature/schema/*.cue by the one generator (charly candy params); nothing hand-written is added beside it.
• R3 — no workaround: no hand edit, no suppression, no retry, no dual shape — the file is the generator's own output and the gate is the same command CI runs.
• R4 — product before prose: no documentation is changed to route around anything; the artifact is corrected at its source (schema → generated file).
• R5 — legacy deleted completely: the stale shape is replaced wholesale, with no shim and no compatibility path left behind.
• R6 — git status was read first; the branch already stood at this exact SHA on origin, so nothing was pushed, amended, rewritten or force-pushed; main is untouched and no other session's work is edited.
• R7 — the check FAILS without the change and passes with it (rc=1 at the base, rc=0 at the head — both pasted), and the emitted artifact is validated at its actual boundary: the same generator that writes the file is the one that judges it.
• R8 — N/A — no live service boundary is crossed: the change is a generated Go type projection, and nothing here is mocked, stubbed or faked.
• R9 — N/A — no binary is rebuilt and no runtime OS dependency is added; the diff touches only the generated projection (and, where noted, a test or the candy's own plan:).
• R10 — not run, and not claimed: no disposable: true bed was run in this round. For a generated-projection regeneration the owning gate is the drift check pasted above; this body claims syntax check only, not a live validation.

Change classification

Generated artifact; no hand edits. candy/plugin-feature/params/cue_types_gen.go is produced by charly candy params plugin-feature from candy/plugin-feature/schema/*.cue, and this PR commits the generator's current output. There is no schema
change, no capability change, no label change and no runtime dependency.

Rollback: git revert the commit on this branch — the previous projection returns and the drift gate goes back
to failing, which is precisely the state main is in today. Nothing to migrate, no data, no consumers to coordinate.

Assisted-by: DeepSeek Harness deepseek-v4.1-flash (syntax check only)

[github-actions]

Review — PASS

Scope & correctness. The PR is a single-file change to candy/plugin-feature/params/cue_types_gen.go (+18/−3), and the diff is exactly that: comment-only edits to the generated FeaturePlugin type. The struct shape (Command, Contract with their yaml/json tags) is untouched, so no runtime behavior, capability, or API surface changes. This matches the body's claim of a pure generated-projection regeneration.

Body honesty. The body's stat block (1 file changed, 18 insertions(+), 3 deletions(-)) reconciles exactly with the diff header (+18/−3) and a line-by-line count of the added comment block (15 + 1 + 2 = 18) against the 3 removed lines. No contradiction between prose and diff.

Generated-artifact claim. The edits are doc comments replacing older text with a longer schema-style block, which is consistent with cue exp gengotypes carrying revised CUE doc comments into the Go output. The content is internally coherent: it documents FeaturePlugin as the plugin's self-contained CUE schema and describes the command:feature pass-through CLI grammar (list/pending/validate) — consistent with the Command/Contract fields it annotates. I cannot re-run the pinned generator or diff against the schema (neither is in the provided state), but the diff's pure-comment nature and internal consistency support the "no hand edits" claim rather than contradicting it. I note this as a verification limitation, not a defect.

Tests. This is a generated projection; the owning gate is the drift check (charly candy params … --check), which the body reports rc=0 at head and rc=1 at base (the defect being removed). For a comment-only regeneration, no additional unit coverage is expected or warranted.

Security. No secrets, no input handling, no dependency or privilege changes — comments only.

Scope & cleanliness. Single coherent file, no shim, no duplicated hand-written path, no dead code introduced.

Comments. The comment thread is empty (0 comments); nothing to disposition.

Rulebook. The R0–R10 self-report is internally consistent; R10 explicitly states "not run, and not claimed" and labels the artifact "syntax check only" — honest, not overclaiming. Rollback (git revert) is accurate for a generated file.

No blocking findings. The change stands on its own evidence: a faithful, minimal regeneration with a consistent body and a matching negative/positive gate claim.

Verdict: PASS

---

ollama-cloud/deepseek-v4.1-flash — action-review.

View action run

[github-actions]

Review — PASS

Scope & correctness

The PR changes exactly one file, candy/plugin-feature/params/cue_types_gen.go (+18/−3). Reading the diff line by line, every added and removed line is a // doc comment; the struct definition (FeaturePlugin), its two fields (Command, Contract), and their yaml/json tags are untouched. No executable code, API surface, capability, label, or dependency changes. This is consistent with the body's claim of a pure generated-projection regeneration.

Body honesty

• The stat block (1 file changed, 18 insertions(+), 3 deletions(-)) reconciles exactly with the diff header and with a line-by-line count of the added comment block against the 3 removed lines.
• The commit list (33f5109… fix + 1b740d2d… merge of main) matches the body's narrative of an unamended fix commit plus a merge-only gh pr update-branch. Nothing contradicts the "no rebase / no force-push" claim.
• The body's R10 explicitly states "not run, and not claimed" and labels the artifact "syntax check only" — honest, not overclaiming a live validation.

Generated-artifact claim

The edits replace older doc text with a longer schema-style block describing FeaturePlugin as a self-contained CUE schema and the command:feature pass-through CLI grammar (list/pending/validate), which is coherent with the fields it annotates. I do not have the CUE schema or the pinned generator in the provided state, so I cannot independently confirm "byte for byte" generation. This is a verification limitation, not a demonstrated defect: the change is comment-only, and the owning drift gate is the very command the body reports failing at base (rc=1) and passing at head (rc=0) — a gate that would re-fail if the file were hand-edited and divergent. I therefore treat the claim as internally consistent rather than contradicted.

Tests

For a generated comment-only projection, the owning check is the drift gate, which the body reports as rc=1 at base / rc=0 at head. No additional unit coverage is warranted for a doc-comment regeneration.

Security

Comments only — no secrets, no input handling, no privilege or dependency changes.

Scope & cleanliness

Single coherent file, no shim, no duplicated hand-written path, no dead code introduced. Rollback (git revert) is accurate for a generated file.

Comments

One comment: github-actions[bot] review claiming PASS. It is an automated review, not a maintainer sign-off, so it does not grant approval — the change stands on its own evidence. Its substantive analysis (comment-only diff, matching line counts, generated-artifact consistency) re-derives correctly against the current diff. Its one factual slip — "The comment thread is empty (0 comments)" while it is itself the single comment — is a trivial self-reference error and does not affect the disposition. Its noted verification limitation (cannot re-run the generator) matches my own.

No blocking findings. The change is a faithful, minimal, comment-only regeneration whose body is internally consistent and whose gate claims are plausible and non-contradicted.

Verdict: PASS

---

ollama-cloud/deepseek-v4.1-flash — action-review.

View action run