@voidh4x0r/slopgate 0.1.16 → 0.1.18

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -9,6 +9,27 @@ All notable changes to this project are documented here. The format follows [Kee
9
9
  - Go and Rust grammars.
10
10
  - VS Code diagnostics.
11
11
 
12
+ ## [0.1.18] - 2026-11-22
13
+
14
+ ### Changed
15
+ - README “Auditing AI-written code” now states the scope of the medium threshold: `pass` / `...` / `throw not implemented` stubs are covered; a bare `return None` is `low`, while `return []` / `{}` / `""` / `False` and empty classes are not reported because they are indistinguishable from legitimate default returns.
16
+ - `--strict` on a scan/rules/MCP command prints a stderr hint that it only affects `init` and points to `--confidence medium`; behavior is otherwise unchanged.
17
+
18
+ ### Tests
19
+ - 272 tests, including the `--strict` misuse warning.
20
+ - Benchmark unchanged at **29 high-confidence findings across 991,705 lines of CPython, 96.6% labeled precision (28/29), 14/14 seeded defects, 0 high findings on 601 JavaScript files**; `staleLabels: []`, `unlabeledHigh: []`.
21
+
22
+ ## [0.1.17] - 2026-11-22
23
+
24
+ ### Added
25
+ - `slopgate init --strict` writes `confidence: "medium"` and `failOn: "medium"` into `.slopgate.json` and generates a workflow with `--confidence medium --fail-on medium`, matching the AI-code audit recipe. Default `init` remains high/high for human-maintained repositories.
26
+ - README “Auditing AI-written code” section with the baseline-first medium recipe, explaining that implementation stubs are medium by design while test theatre is high.
27
+ - `py/stub-body` and `ts/stub-function` docs now state that most real AI-stub findings are at medium confidence.
28
+
29
+ ### Changed
30
+ - 271 tests, including `init --strict` config and workflow assertions.
31
+ - Benchmark unchanged at **29 high-confidence findings across 991,705 lines of CPython, 96.6% labeled precision (28/29), 14/14 seeded defects, 0 high findings on 601 JavaScript files**; `staleLabels: []`, `unlabeledHigh: []`.
32
+
12
33
  ## [0.1.16] - 2026-11-22
13
34
 
14
35
  ### Fixed
package/README.md CHANGED
@@ -14,7 +14,7 @@ AI writes code that *looks* right. `slopgate` finds the parts that only look rig
14
14
  ```console
15
15
  $ npx @voidh4x0r/slopgate .
16
16
 
17
- slopgate 0.1.16 · 4 files · 39ms
17
+ slopgate 0.1.18 · 4 files · 39ms
18
18
 
19
19
  app.py
20
20
  4:1 error high py/phantom-import `helper` is imported but not declared in any manifest and not present in this repository
@@ -172,6 +172,22 @@ npx @voidh4x0r/slopgate . --fail-on high
172
172
  npx @voidh4x0r/slopgate . --diff origin/main --fail-on high
173
173
  ```
174
174
 
175
+ ### Auditing AI-written code
176
+
177
+ The defaults are tuned for human-maintained code: an interface stub in a library is often deliberate, so class-method stubs and documented `pass` bodies are `medium`. AI-generated code has the opposite prior — an unimplemented method is usually unfinished — so audit it with the medium threshold:
178
+
179
+ ```bash
180
+ # 1. Absorb the existing medium-confidence debt once.
181
+ npx @voidh4x0r/slopgate . --confidence medium --fail-on medium --write-baseline
182
+
183
+ # 2. Gate NEW implementation stubs and test theatre at medium.
184
+ npx @voidh4x0r/slopgate . --confidence medium --fail-on medium
185
+ ```
186
+
187
+ `slopgate init --strict` writes this policy into `.slopgate.json` and the generated GitHub Actions workflow. Keep the default `init` for human-maintained repositories.
188
+
189
+ **Scope of medium:** the medium threshold covers stubs whose entire body is `pass`, `...` or `throw new Error("not implemented")`. A bare `return None` is reported at `low`; `return []`, `{}`, `""`, `False` and empty classes are **not reported**, because they are indistinguishable from legitimate default return values. Do not expect strict mode to catch those placeholder returns.
190
+
175
191
  ### Use it from a coding agent (MCP)
176
192
 
177
193
  `slopgate` ships a Model Context Protocol server over stdio, so Claude Code, Codex and other MCP clients can check their own diff **before** saying "done".
@@ -210,7 +226,7 @@ jobs:
210
226
  - uses: actions/checkout@v4
211
227
  - uses: actions/setup-node@v4
212
228
  with: { node-version: 22 }
213
- - run: npx --yes @voidh4x0r/slopgate@0.1.16 . --fail-on high
229
+ - run: npx --yes @voidh4x0r/slopgate@0.1.18 . --fail-on high
214
230
  ```
215
231
 
216
232
  For GitHub code scanning, use `--format sarif` and upload with `github/codeql-action/upload-sarif`.
@@ -218,7 +234,7 @@ For GitHub code scanning, use `--format sarif` and upload with `github/codeql-ac
218
234
  For inline pull-request annotations **without** code scanning, use `--format github`, which emits [workflow commands](https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions):
219
235
 
220
236
  ```yaml
221
- - run: npx --yes @voidh4x0r/slopgate@0.1.16 . --format github --fail-on high
237
+ - run: npx --yes @voidh4x0r/slopgate@0.1.18 . --format github --fail-on high
222
238
  ```
223
239
 
224
240
  Each finding becomes an `error` / `warning` / `notice` annotation on the exact line, followed by a summary notice.
@@ -303,7 +319,7 @@ Issues and pull requests are welcome. Start with [CONTRIBUTING.md](https://githu
303
319
  git clone https://github.com/VoidH4x0r/slopgate
304
320
  cd slopgate
305
321
  npm install
306
- npm test # 270 tests; the tests themselves make no network or child-process calls
322
+ npm test # 272 tests; the tests themselves make no network or child-process calls
307
323
  npm run typecheck
308
324
  npm run verify # all release gates in one command
309
325
  ```
package/dist/cli.js CHANGED
@@ -24,7 +24,7 @@ Usage:
24
24
  slopgate [path] [options] scan a directory or file (default: .)
25
25
  slopgate rules list every rule
26
26
  slopgate mcp [path] run as an MCP server on stdio
27
- slopgate init add config + GitHub Action
27
+ slopgate init [--strict] add config + GitHub Action
28
28
  slopgate --version | --help
29
29
 
30
30
  Options:
@@ -41,6 +41,7 @@ Options:
41
41
  --quiet only print the summary
42
42
  --verbose print warnings and per-file details
43
43
  --force overwrite files in \`init\`
44
+ --strict \`init\`: medium confidence + fail-on medium
44
45
  -h, --help show this help
45
46
  -V, --version print the version
46
47
 
@@ -77,6 +78,7 @@ function defaultOptions() {
77
78
  quiet: false,
78
79
  verbose: false,
79
80
  force: false,
81
+ strict: false,
80
82
  };
81
83
  }
82
84
  const VALUE_FLAGS = new Set([
@@ -137,6 +139,10 @@ function parseArgs(argv) {
137
139
  options.force = true;
138
140
  continue;
139
141
  }
142
+ if (token === '--strict') {
143
+ options.strict = true;
144
+ continue;
145
+ }
140
146
  const equals = token.indexOf('=');
141
147
  const rawName = (equals === -1 ? token : token.slice(0, equals)).replace(/^--?/, '');
142
148
  if (!VALUE_FLAGS.has(rawName))
@@ -265,7 +271,10 @@ const INIT_CONFIG = {
265
271
  confidence: 'high',
266
272
  failOn: 'high',
267
273
  };
268
- const INIT_WORKFLOW = (version, target) => `name: slopgate
274
+ function initConfig(strict) {
275
+ return strict ? { ...INIT_CONFIG, confidence: 'medium', failOn: 'medium' } : INIT_CONFIG;
276
+ }
277
+ const INIT_WORKFLOW = (version, target, strict) => `name: slopgate
269
278
 
270
279
  on:
271
280
  push:
@@ -285,7 +294,7 @@ jobs:
285
294
  with:
286
295
  node-version: 22
287
296
  - name: Scan for AI slop
288
- run: npx --yes @voidh4x0r/slopgate@${version} ${target} --fail-on high --format pretty
297
+ run: npx --yes @voidh4x0r/slopgate@${version} ${target} ${strict ? '--confidence medium --fail-on medium' : '--fail-on high'} --format pretty
289
298
  `;
290
299
  async function runInit(options, io, version) {
291
300
  const requested = path.resolve(options.paths[0] ?? '.');
@@ -303,7 +312,7 @@ async function runInit(options, io, version) {
303
312
  warn(io, options, `${path.relative(process.cwd(), configPath)} already exists (use --force to overwrite)`);
304
313
  }
305
314
  else {
306
- await writeFile(configPath, JSON.stringify(INIT_CONFIG, null, 2) + '\n', 'utf8');
315
+ await writeFile(configPath, JSON.stringify(initConfig(options.strict), null, 2) + '\n', 'utf8');
307
316
  writeOut(io, `created ${path.relative(process.cwd(), configPath)}`);
308
317
  }
309
318
  if (!options.force && (await exists(workflowPath))) {
@@ -311,10 +320,13 @@ async function runInit(options, io, version) {
311
320
  }
312
321
  else {
313
322
  await mkdir(path.dirname(workflowPath), { recursive: true });
314
- await writeFile(workflowPath, INIT_WORKFLOW(version, workflowTarget), 'utf8');
323
+ await writeFile(workflowPath, INIT_WORKFLOW(version, workflowTarget, options.strict), 'utf8');
315
324
  writeOut(io, `created ${path.relative(process.cwd(), workflowPath)}`);
316
325
  }
317
- writeOut(io, `next: npx @voidh4x0r/slopgate ${workflowTarget} --write-baseline, then commit the baseline`);
326
+ const policy = options.strict
327
+ ? '--confidence medium --fail-on medium'
328
+ : '--fail-on high';
329
+ writeOut(io, `next: npx @voidh4x0r/slopgate ${workflowTarget} ${policy} --write-baseline, then commit the baseline`);
318
330
  return 0;
319
331
  }
320
332
  async function runMcp(options) {
@@ -473,6 +485,9 @@ export async function main(argv, io = defaultIo()) {
473
485
  io.err(`slopgate: ${error}\n\n${HELP}`);
474
486
  return 2;
475
487
  }
488
+ if (options.strict && options.command !== 'init' && options.command !== 'help' && options.command !== 'version') {
489
+ warn(io, options, '--strict only affects `init`; use `--confidence medium` for scans');
490
+ }
476
491
  try {
477
492
  switch (options.command) {
478
493
  case 'help':
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@voidh4x0r/slopgate",
3
- "version": "0.1.16",
3
+ "version": "0.1.18",
4
4
  "description": "The slop gate for AI-written code. Precision-first structural linting for Python and TypeScript.",
5
5
  "type": "module",
6
6
  "license": "MIT",