ql-deslop

Name: ql-deslop
Author: andyzengmath

23stars0forks

How to install ql-deslop?

andyzengmath/quantum-loop/ql-deslop

$npx -y skills add andyzengmath/quantum-loop --skill ql-deslop

Installs into the current project.

Summary

Post-review AI-slop cleanup pass. After quality review approves a story/feature, scans only the changed files for: duplicate code, dead code, needless abstraction, boundary violations, missing tests. Proposes deletions with regression-test gate — rolls back on regression. Mandato

Files · 1

View on GitHub

SKILL.md

1	# ql-deslop — post-review AI-slop cleanup
2
3	## Purpose
4
5	Review approval does not mean the code is clean. LLM-authored code routinely introduces:
6	- Duplication of helpers / types / imports under slightly different names.
7	- Dead code — functions added but never called; exports that no consumer uses.
8	- Needless abstraction — factory for one caller; wrapper that only aliases; configurable field never varied.
9	- Boundary violations — private functions accessed via string names; cross-module reaches that skip the public API.
10	- Missing tests — paths and branches not exercised by any test.
11
12	Per Schreiber & Tippe 2025, 12.1% of AI-generated files contain ≥1 CWE, and per Zhong et al. 2026 (278k conversations) AI suggestions increase cyclomatic complexity 10-50× more than human edits. Without an explicit cleanup step, every autonomous run accumulates slop.
13
14	`ql-deslop` adds an explicit cleanup pass AFTER the story's review gate passes but BEFORE the merge to the feature branch. Borrowed directly from OMC Ralph Steps 7.5 (deslop) + 7.6 (regression re-verify).
15
16	## When to use
17
18	- Automatically, as a mandatory step between `ql-review` PASS and the `git merge` in `ql-execute`.
19	- Manually, at any point after a story is implemented, to clean up before review.
20	- After absorbing a foreign branch (CPC promotion, cherry-pick) — the absorbed work may carry slop.
21
22	## Opt-out
23
24	Single flag: `--no-deslop`. Used only when:
25	- The story is a pure test addition (nothing to clean).
26	- The deslop pass itself was run earlier in the same wave on the same files.
27	- The user explicitly demands skip (rare; recorded in commit trailer `Deslop: skipped \| <reason>`).
28
29	## Scope — strict file-list only
30
31	The deslop pass operates on `git diff --name-only BASE_SHA..HEAD_SHA` — the EXACT files the story touched. It never broadens scope silently. Reaching into unchanged files is an anti-pattern and is explicitly forbidden.
32
33	## Smell taxonomy — 5 categories
34
35	### 1. Duplication
36
37	Detectors:
38	- Identifier similarity across changed files: extract top-level symbol names; flag near-duplicates (e.g., `parseRequest` / `parseReq` / `handleParse`) via Levenshtein or canonical-alpha-rename + exact match (per HyClone arXiv:2508.01357).
39	- Function-body AST hash: for each function in the diff, compute tree-sitter AST hash after alpha-renaming; flag duplicates across files.
40	- Import redundancy: same module imported as different names in sibling files.
41
42	Action: extract common code into a shared lib; update callers; delete the duplicates.
43
44	### 2. Dead code
45
46	Detectors:
47	- Language-specific reachability analysis:
48	- TypeScript/JavaScript: `knip` or `ts-prune` on the changed-files subset.
49	- Python: `vulture` on the changed-files subset.
50	- Go: `staticcheck -checks=U1000` on the changed-files subset.
51	- Rust: `cargo udeps` on the changed-files subset.
52	- DePA line-level perplexity (optional; per arXiv:2502.20246): high-z anomalies flagged for manual inspection.
53
54	Action: delete the unused export / function / variable. Re-run tests.
55
56	### 3. Needless abstraction
57
58	Detectors:
59	- Single-caller factories: factory returns type T, only one call site exists → inline the factory.
60	- Identity wrappers: function that only calls another function with same arguments → inline or delete.
61	- Never-varied config: options object with fields that are always set to the same literal → remove the option.
62	- One-member interface implemented by one class → inline unless specified as a contract.
63
64	Action: propose refactoring with regression-test gate.
65
66	### 4. Boundary violations
67
68	Detectors:
69	- Name mangling: access via `module["_privateFn"]`, reflection-style private access.
70	- Path traversal: imports that reach up too many levels (`../../../../other-module/internal`).
71	- Test imports from production: test files reaching into non-exported implementation details.
72
73	Action: reshape through the public API, or promote the private API if genuinely needed externally.
74
75	### 5. Missing tests
76
77	Detectors:
78	- AC ↔ test mapping: every AC in the PRD should have a specific test referenced by its ID (`AC-3 → tests/feature.test.ts::test_ac_3`). Missing maps surface as findings.
79	- Coverage gap on new public functions: any new exported function with <1 test invocation.
80	- Branch-coverage gap on control-flow-heavy changes.
81
82	Action: write the missing test BEFORE deleting alleged dead code (deletion is safe only when the test suite exercises the surviving behavior).
83
84	## Single-pass discipline
85
86	Each deslop invocation runs ONE pass from the taxonomy at a time:
87
88	- **Pass 1 — dead

People who liked this also viewed

6 picks

Type

ql-deslop

23stars0forks

View on GitHub

How to install ql-deslop?

andyzengmath/quantum-loop/ql-deslop

$npx -y skills add andyzengmath/quantum-loop --skill ql-deslop

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# ql-deslop — post-review AI-slop cleanup
2
3	## Purpose
4
5	Review approval does not mean the code is clean. LLM-authored code routinely introduces:
6	- Duplication of helpers / types / imports under slightly different names.
7	- Dead code — functions added but never called; exports that no consumer uses.
8	- Needless abstraction — factory for one caller; wrapper that only aliases; configurable field never varied.
9	- Boundary violations — private functions accessed via string names; cross-module reaches that skip the public API.
10	- Missing tests — paths and branches not exercised by any test.
11
12	Per Schreiber & Tippe 2025, 12.1% of AI-generated files contain ≥1 CWE, and per Zhong et al. 2026 (278k conversations) AI suggestions increase cyclomatic complexity 10-50× more than human edits. Without an explicit cleanup step, every autonomous run accumulates slop.
13
14	`ql-deslop` adds an explicit cleanup pass AFTER the story's review gate passes but BEFORE the merge to the feature branch. Borrowed directly from OMC Ralph Steps 7.5 (deslop) + 7.6 (regression re-verify).
15
16	## When to use
17
18	- Automatically, as a mandatory step between `ql-review` PASS and the `git merge` in `ql-execute`.
19	- Manually, at any point after a story is implemented, to clean up before review.
20	- After absorbing a foreign branch (CPC promotion, cherry-pick) — the absorbed work may carry slop.
21
22	## Opt-out
23
24	Single flag: `--no-deslop`. Used only when:
25	- The story is a pure test addition (nothing to clean).
26	- The deslop pass itself was run earlier in the same wave on the same files.
27	- The user explicitly demands skip (rare; recorded in commit trailer `Deslop: skipped \| <reason>`).
28
29	## Scope — strict file-list only
30
31	The deslop pass operates on `git diff --name-only BASE_SHA..HEAD_SHA` — the EXACT files the story touched. It never broadens scope silently. Reaching into unchanged files is an anti-pattern and is explicitly forbidden.
32
33	## Smell taxonomy — 5 categories
34
35	### 1. Duplication
36
37	Detectors:
38	- Identifier similarity across changed files: extract top-level symbol names; flag near-duplicates (e.g., `parseRequest` / `parseReq` / `handleParse`) via Levenshtein or canonical-alpha-rename + exact match (per HyClone arXiv:2508.01357).
39	- Function-body AST hash: for each function in the diff, compute tree-sitter AST hash after alpha-renaming; flag duplicates across files.
40	- Import redundancy: same module imported as different names in sibling files.
41
42	Action: extract common code into a shared lib; update callers; delete the duplicates.
43
44	### 2. Dead code
45
46	Detectors:
47	- Language-specific reachability analysis:
48	- TypeScript/JavaScript: `knip` or `ts-prune` on the changed-files subset.
49	- Python: `vulture` on the changed-files subset.
50	- Go: `staticcheck -checks=U1000` on the changed-files subset.
51	- Rust: `cargo udeps` on the changed-files subset.
52	- DePA line-level perplexity (optional; per arXiv:2502.20246): high-z anomalies flagged for manual inspection.
53
54	Action: delete the unused export / function / variable. Re-run tests.
55
56	### 3. Needless abstraction
57
58	Detectors:
59	- Single-caller factories: factory returns type T, only one call site exists → inline the factory.
60	- Identity wrappers: function that only calls another function with same arguments → inline or delete.
61	- Never-varied config: options object with fields that are always set to the same literal → remove the option.
62	- One-member interface implemented by one class → inline unless specified as a contract.
63
64	Action: propose refactoring with regression-test gate.
65
66	### 4. Boundary violations
67
68	Detectors:
69	- Name mangling: access via `module["_privateFn"]`, reflection-style private access.
70	- Path traversal: imports that reach up too many levels (`../../../../other-module/internal`).
71	- Test imports from production: test files reaching into non-exported implementation details.
72
73	Action: reshape through the public API, or promote the private API if genuinely needed externally.
74
75	### 5. Missing tests
76
77	Detectors:
78	- AC ↔ test mapping: every AC in the PRD should have a specific test referenced by its ID (`AC-3 → tests/feature.test.ts::test_ac_3`). Missing maps surface as findings.
79	- Coverage gap on new public functions: any new exported function with <1 test invocation.
80	- Branch-coverage gap on control-flow-heavy changes.
81
82	Action: write the missing test BEFORE deleting alleged dead code (deletion is safe only when the test suite exercises the surviving behavior).
83
84	## Single-pass discipline
85
86	Each deslop invocation runs ONE pass from the taxonomy at a time:
87
88	- **Pass 1 — dead

People who liked this also viewed

6 picks

Type