jeffy-loop 1.26.6__tar.gz → 1.26.7__tar.gz

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.
Files changed (24) hide show
  1. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/PKG-INFO +8 -4
  2. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/README.md +7 -3
  3. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/pyproject.toml +1 -1
  4. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/stop-hook.sh +1 -1
  5. jeffy_loop-1.26.7/src/jeffy_loop/__init__.py +1 -0
  6. jeffy_loop-1.26.6/src/jeffy_loop/__init__.py +0 -1
  7. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/.gitignore +0 -0
  8. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/LICENSE +0 -0
  9. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/hatch_build.py +0 -0
  10. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/cancel-jeffy/SKILL.md +0 -0
  11. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/SKILL.md +0 -0
  12. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/lib/check-claims.sh +0 -0
  13. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/lib/detect-sandbox.sh +0 -0
  14. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/lib/measure-surface.sh +0 -0
  15. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/lib/quiet-verify.sh +0 -0
  16. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/hooks/lib/run-probe.sh +0 -0
  17. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/backlog-default.md +0 -0
  18. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/backlog-highs.md +0 -0
  19. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/iteration-prompt-highs.txt +0 -0
  20. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/iteration-prompt.txt +0 -0
  21. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/journal-default.md +0 -0
  22. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/plan-default.md +0 -0
  23. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/skills/jeffy/references/plan-highs.md +0 -0
  24. {jeffy_loop-1.26.6 → jeffy_loop-1.26.7}/src/jeffy_loop/cli.py +0 -0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.5
2
2
  Name: jeffy-loop
3
- Version: 1.26.6
3
+ Version: 1.26.7
4
4
  Summary: Jeffy Loop is an autonomous software-engineering control system that forces an LLM to continuously audit, act, verify, checkpoint, and prove convergence.
5
5
  Project-URL: Repository, https://github.com/lenamonj/jeffy-loop
6
6
  License-Expression: MIT
@@ -25,7 +25,7 @@ Description-Content-Type: text/markdown
25
25
 
26
26
  **[Quick Install](#quick-install)**  ·  **[Usage](https://github.com/lenamonj/jeffy-loop/blob/main/docs/usage.md)**  ·  **[How it works](https://github.com/lenamonj/jeffy-loop/blob/main/docs/how-it-works.md)**  ·  **[The receipts](https://github.com/lenamonj/jeffy-loop/blob/main/evals/README.md)**  ·  **[Headless](https://github.com/lenamonj/jeffy-loop/blob/main/docs/headless.md)**  ·  **[White paper](https://github.com/lenamonj/jeffy-loop/raw/main/The-Jeffy-Loop.pdf)**  ·  **[STE100 description](https://github.com/lenamonj/jeffy-loop/raw/main/Jeffy-Loop-STE100.pdf)**
27
27
 
28
- ## A Claude Code loop that finds bugs in your repo, fixes them, and proves every fix before it commits. Maintainers have merged <!-- count:merged -->70<!-- /count --> of its patches.
28
+ ## A Claude Code loop that finds bugs in your repo, fixes them, and proves every fix before it commits. Maintainers have merged <!-- count:merged -->71<!-- /count --> of its patches.
29
29
 
30
30
  </div>
31
31
 
@@ -33,7 +33,7 @@ The model writes the patch. The harness around it decides whether that patch can
33
33
 
34
34
  The harness is deliberately hard to fool. If a fix breaks your tests, it is undone. Nothing is ever pushed. A standard run counts as complete only when a fresh audit is clean, your tests pass again, and an adversarial review by a second AI agent with no part in the run signs off. The goal is fixes that survive scrutiny.
35
35
 
36
- There is a real-world test for that. Maintainers with no stake in this project have merged <!-- count:merged -->70<!-- /count --> of its pull requests, each filed from a local clone, into projects run by NVIDIA, Arm, Meta, Tesla, Google, Apple, Microsoft, Netflix, Apache, Palo Alto Networks, Oracle, IBM, HashiCorp, Cisco, Uber, Square, Cloudflare, and more. [See them all.](#independent-validation---the-receipts)
36
+ There is a real-world test for that. Maintainers with no stake in this project have merged <!-- count:merged -->71<!-- /count --> of its pull requests, each filed from a local clone, into projects run by NVIDIA, Arm, Meta, Tesla, Google, Apple, Microsoft, Netflix, Apache, Palo Alto Networks, Oracle, IBM, HashiCorp, Cisco, Uber, Square, Cloudflare, Dropbox, and more. [See them all.](#independent-validation---the-receipts)
37
37
 
38
38
  <div align="center">
39
39
 
@@ -50,7 +50,7 @@ Jeffy was run against <!-- count:tested -->132<!-- /count --> open-source projec
50
50
 
51
51
  | Projects tested | Converged | Failed | PRs merged | PRs open | Issues filed |
52
52
  |:---:|:---:|:---:|:---:|:---:|:---:|
53
- | **<!-- count:tested -->132<!-- /count -->** | **<!-- count:converged -->103<!-- /count -->** | **<!-- count:failed -->28<!-- /count -->** | **<!-- count:merged -->70<!-- /count -->** | **<!-- count:prs-open -->17<!-- /count -->** | **<!-- count:issues -->4<!-- /count -->** |
53
+ | **<!-- count:tested -->132<!-- /count -->** | **<!-- count:converged -->103<!-- /count -->** | **<!-- count:failed -->28<!-- /count -->** | **<!-- count:merged -->71<!-- /count -->** | **<!-- count:prs-open -->17<!-- /count -->** | **<!-- count:issues -->4<!-- /count -->** |
54
54
 
55
55
  **Converged:** the closing audit came back clean and the loop's adversarial evaluator, a fresh-context sub-agent, countersigned it, a standard this repository set and checks itself. That happened in <!-- count:converged -->103<!-- /count --> projects across <!-- count:languages -->13<!-- /count --> languages with no language-specific analyzer. **Failed:** the run spent the budget declared before it started without converging, or, for libuv, was abandoned before it had one. That leaves PapaParse, an audit held to the same method rather than a loop run, which the receipts page counts as Fixed alongside the 103.
56
56
 
@@ -251,6 +251,10 @@ Each finding below was accepted upstream by the project's own maintainers.
251
251
  <td><img src="https://github.com/canonical.png" width="20" height="20" alt="" align="absmiddle"> Canonical (Ubuntu)</td>
252
252
  <td><a href="https://github.com/canonical/pebble/pull/938">pebble #938</a> - merged in 16 days<br>With more than one service or check, the metrics endpoint repeated each <code># HELP</code> and <code># TYPE</code> family and the Prometheus parser rejected the whole response<br><br><a href="https://github.com/canonical/pebble/pull/937">pebble #937</a> - merged in 43 days<br>Appending a layer that parsed but failed plan validation left the daemon's live layer list changed, because <code>slices.Insert</code> shifted the existing slice in place before validation ran; the insert now goes into a copy</td>
253
253
  </tr>
254
+ <tr>
255
+ <td><img src="https://github.com/dropbox.png" width="20" height="20" alt="" align="absmiddle"> Dropbox</td>
256
+ <td><a href="https://github.com/dropbox/rust-brotli/pull/267">rust-brotli #267</a> - merged in 48 days<br>Under the <code>disallow_large_window_size</code> feature the distance histogram was sized 520 while every bound still read 544, so compressing ordinary text panicked with an index out of bounds</td>
257
+ </tr>
254
258
  <tr>
255
259
  <td><img src="https://github.com/SethMMorton.png" width="20" height="20" alt="" align="absmiddle"> natsort</td>
256
260
  <td><a href="https://github.com/SethMMorton/natsort/pull/196">natsort #196</a> - merged in 8 days<br>The locale sentinel meant to sort last was three ASCII bytes, so PyICU keys sorted after it (19 million PyPI downloads a month)</td>
@@ -13,7 +13,7 @@
13
13
 
14
14
  **[Quick Install](#quick-install)** &nbsp;·&nbsp; **[Usage](docs/usage.md)** &nbsp;·&nbsp; **[How it works](docs/how-it-works.md)** &nbsp;·&nbsp; **[The receipts](evals/README.md)** &nbsp;·&nbsp; **[Headless](docs/headless.md)** &nbsp;·&nbsp; **[White paper](https://github.com/lenamonj/jeffy-loop/raw/main/The-Jeffy-Loop.pdf)** &nbsp;·&nbsp; **[STE100 description](https://github.com/lenamonj/jeffy-loop/raw/main/Jeffy-Loop-STE100.pdf)**
15
15
 
16
- ## A Claude Code loop that finds bugs in your repo, fixes them, and proves every fix before it commits. Maintainers have merged <!-- count:merged -->70<!-- /count --> of its patches.
16
+ ## A Claude Code loop that finds bugs in your repo, fixes them, and proves every fix before it commits. Maintainers have merged <!-- count:merged -->71<!-- /count --> of its patches.
17
17
 
18
18
  </div>
19
19
 
@@ -21,7 +21,7 @@ The model writes the patch. The harness around it decides whether that patch can
21
21
 
22
22
  The harness is deliberately hard to fool. If a fix breaks your tests, it is undone. Nothing is ever pushed. A standard run counts as complete only when a fresh audit is clean, your tests pass again, and an adversarial review by a second AI agent with no part in the run signs off. The goal is fixes that survive scrutiny.
23
23
 
24
- There is a real-world test for that. Maintainers with no stake in this project have merged <!-- count:merged -->70<!-- /count --> of its pull requests, each filed from a local clone, into projects run by NVIDIA, Arm, Meta, Tesla, Google, Apple, Microsoft, Netflix, Apache, Palo Alto Networks, Oracle, IBM, HashiCorp, Cisco, Uber, Square, Cloudflare, and more. [See them all.](#independent-validation---the-receipts)
24
+ There is a real-world test for that. Maintainers with no stake in this project have merged <!-- count:merged -->71<!-- /count --> of its pull requests, each filed from a local clone, into projects run by NVIDIA, Arm, Meta, Tesla, Google, Apple, Microsoft, Netflix, Apache, Palo Alto Networks, Oracle, IBM, HashiCorp, Cisco, Uber, Square, Cloudflare, Dropbox, and more. [See them all.](#independent-validation---the-receipts)
25
25
 
26
26
  <div align="center">
27
27
 
@@ -38,7 +38,7 @@ Jeffy was run against <!-- count:tested -->132<!-- /count --> open-source projec
38
38
 
39
39
  | Projects tested | Converged | Failed | PRs merged | PRs open | Issues filed |
40
40
  |:---:|:---:|:---:|:---:|:---:|:---:|
41
- | **<!-- count:tested -->132<!-- /count -->** | **<!-- count:converged -->103<!-- /count -->** | **<!-- count:failed -->28<!-- /count -->** | **<!-- count:merged -->70<!-- /count -->** | **<!-- count:prs-open -->17<!-- /count -->** | **<!-- count:issues -->4<!-- /count -->** |
41
+ | **<!-- count:tested -->132<!-- /count -->** | **<!-- count:converged -->103<!-- /count -->** | **<!-- count:failed -->28<!-- /count -->** | **<!-- count:merged -->71<!-- /count -->** | **<!-- count:prs-open -->17<!-- /count -->** | **<!-- count:issues -->4<!-- /count -->** |
42
42
 
43
43
  **Converged:** the closing audit came back clean and the loop's adversarial evaluator, a fresh-context sub-agent, countersigned it, a standard this repository set and checks itself. That happened in <!-- count:converged -->103<!-- /count --> projects across <!-- count:languages -->13<!-- /count --> languages with no language-specific analyzer. **Failed:** the run spent the budget declared before it started without converging, or, for libuv, was abandoned before it had one. That leaves PapaParse, an audit held to the same method rather than a loop run, which the receipts page counts as Fixed alongside the 103.
44
44
 
@@ -239,6 +239,10 @@ Each finding below was accepted upstream by the project's own maintainers.
239
239
  <td><img src="https://github.com/canonical.png" width="20" height="20" alt="" align="absmiddle"> Canonical (Ubuntu)</td>
240
240
  <td><a href="https://github.com/canonical/pebble/pull/938">pebble #938</a> - merged in 16 days<br>With more than one service or check, the metrics endpoint repeated each <code># HELP</code> and <code># TYPE</code> family and the Prometheus parser rejected the whole response<br><br><a href="https://github.com/canonical/pebble/pull/937">pebble #937</a> - merged in 43 days<br>Appending a layer that parsed but failed plan validation left the daemon's live layer list changed, because <code>slices.Insert</code> shifted the existing slice in place before validation ran; the insert now goes into a copy</td>
241
241
  </tr>
242
+ <tr>
243
+ <td><img src="https://github.com/dropbox.png" width="20" height="20" alt="" align="absmiddle"> Dropbox</td>
244
+ <td><a href="https://github.com/dropbox/rust-brotli/pull/267">rust-brotli #267</a> - merged in 48 days<br>Under the <code>disallow_large_window_size</code> feature the distance histogram was sized 520 while every bound still read 544, so compressing ordinary text panicked with an index out of bounds</td>
245
+ </tr>
242
246
  <tr>
243
247
  <td><img src="https://github.com/SethMMorton.png" width="20" height="20" alt="" align="absmiddle"> natsort</td>
244
248
  <td><a href="https://github.com/SethMMorton/natsort/pull/196">natsort #196</a> - merged in 8 days<br>The locale sentinel meant to sort last was three ASCII bytes, so PyICU keys sorted after it (19 million PyPI downloads a month)</td>
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
4
4
 
5
5
  [project]
6
6
  name = "jeffy-loop"
7
- version = "1.26.6"
7
+ version = "1.26.7"
8
8
  description = "Jeffy Loop is an autonomous software-engineering control system that forces an LLM to continuously audit, act, verify, checkpoint, and prove convergence."
9
9
  dynamic = ["readme"]
10
10
  license = "MIT"
@@ -10,7 +10,7 @@
10
10
  # directory Claude Code was started in, so Bash-tool cwd drift mid-iteration
11
11
  # cannot kill the loop.
12
12
  set -u
13
- JEFFY_VERSION="1.26.6"
13
+ JEFFY_VERSION="1.26.7"
14
14
 
15
15
  # 1.21.0 (P1-70): `stop-hook.sh --lint [project root]` derives every check
16
16
  # the declaration path enforces and prints the first refusal, or clean,
@@ -0,0 +1 @@
1
+ __version__ = "1.26.7"
@@ -1 +0,0 @@
1
- __version__ = "1.26.6"
File without changes
File without changes
File without changes