@shadowclone/cli 0.0.16 → 0.0.17

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 (23) hide show
  1. package/PRIVACY.md +1 -1
  2. package/README.md +32 -29
  3. package/dist/_.._/web/client/index.html +24 -40
  4. package/dist/shadowclone.js +1158 -508
  5. package/dist/web/{index-dzztba9n.css → index-ntyv9x8a.css} +1 -1
  6. package/dist/web/index-twq4tm3n.js +135 -0
  7. package/evals.md +117 -267
  8. package/package.json +9 -3
  9. package/plugins/shadowclone/.claude-plugin/plugin.json +1 -1
  10. package/plugins/shadowclone/.codex-plugin/plugin.json +1 -1
  11. package/plugins/shadowclone/plugin.json +1 -1
  12. package/skills/design-deep-modules/SKILL.md +2 -0
  13. package/skills/diagnose-before-editing/SKILL.md +2 -0
  14. package/skills/prove-regression-tests/SKILL.md +2 -0
  15. package/skills/research-primary-sources/SKILL.md +2 -0
  16. package/skills/resolve-conflicts-by-intent/SKILL.md +2 -0
  17. package/skills/scope-confirmed-changes/SKILL.md +2 -0
  18. package/skills/shadowclone-work/SKILL.md +26 -20
  19. package/skills/testing-first/SKILL.md +2 -0
  20. package/skills/testing-risk-based/SKILL.md +2 -0
  21. package/skills/typescript-type-safety/SKILL.md +2 -0
  22. package/skills/verify-and-review/SKILL.md +2 -0
  23. package/dist/web/index-s1e5043w.js +0 -134
package/PRIVACY.md CHANGED
@@ -10,7 +10,7 @@ Shadowclone excludes tool results, tool-returned file contents, thinking blocks,
10
10
 
11
11
  ## What leaves your machine
12
12
 
13
- Opening the browser editor and using the local MCP server do not send data to Shadowclone. Optional learning, skill drafting, build descriptions, and delegated runs use an agent CLI you choose. Consented, redacted inputs go directly through that provider and are governed by its terms.
13
+ Opening the browser editor and using the local MCP server do not send data to Shadowclone. Optional learning, skill drafting, and build descriptions use an agent CLI you choose. Consented, redacted inputs go directly through that provider and are governed by its terms.
14
14
 
15
15
  The build description action sends selected skill titles and redacted summaries. It does not send skill bodies, file paths, repository names, or ownership metadata.
16
16
 
package/README.md CHANGED
@@ -34,11 +34,11 @@ Install the CLI:
34
34
  npm install -g @shadowclone/cli
35
35
  ```
36
36
 
37
- | Step | Command | What happens | Time |
38
- | --- | --- | --- | ---: |
39
- | 1 | `shadowclone init` | Answer three questions about learning and skill maintenance | About 1 minute |
40
- | 2 | `shadowclone learn --deep` | Learn from enabled past sessions; repeat if more history remains | Up to 5 minutes per run |
41
- | 3 | `shadowclone learning pending` and `shadowclone skills pending` | Review learned rules, scope, and publication before relying on them | About 1 minute |
37
+ | Step | Command | What happens | Time |
38
+ | ---- | --------------------------------------------------------------- | ------------------------------------------------------------------- | ----------------------: |
39
+ | 1 | `shadowclone init` | Answer three questions about learning and skill maintenance | About 1 minute |
40
+ | 2 | `shadowclone learn --deep` | Learn from enabled past sessions; repeat if more history remains | Up to 5 minutes per run |
41
+ | 3 | `shadowclone learning pending` and `shadowclone skills pending` | Review learned rules, scope, and publication before relying on them | About 1 minute |
42
42
 
43
43
  Setup takes about two to three minutes, excluding learning. Deep learning needs an installed and configured `claude`, `codex`, `cursor-agent`, or `pi` CLI. [Pi setup](docs/guides/pi.md) uses any suitable model configured in Pi, including local providers. Claude Code learning on Linux needs `bwrap` and `socat` for its sandbox.
44
44
 
@@ -46,22 +46,25 @@ Setup takes about two to three minutes, excluding learning. Deep learning needs
46
46
 
47
47
  Run `shadowclone` to see the current result and next action. Use `shadowclone wizard` to browse optional workflow skills in a browser.
48
48
 
49
- **Delegate a task.** The optional `shadowclone-work` skill records your standards, acceptance criteria, checks, and review in a private task receipt. It adds work tracking to an existing Claude Code or Codex session. Start with review handoff; shipping actions require explicit repository grants. Native workflow qualification is still pending. [Use delegated work](docs/guides/delegated-work.md).
49
+ **Delegate a pull request.** The optional `shadowclone-work` skill takes a request, an issue, or an existing PR to ready for review: green checks, conflicts resolved, and review comments fixed or left for you. It replies to fixed comments with commit SHAs only and never merges. [Use delegated work](docs/guides/delegated-work.md).
50
+
51
+ **Use a personal GitHub clone.** Run `shadowclone bot setup` from a repository checkout. Name your App, select its repositories, and review the exact guidance before connecting your Claude subscription. The clone handles owner issues and tagged PR requests. Live installation qualification is pending. [Set up a GitHub clone](docs/guides/github-clones.md).
50
52
 
51
53
  ## Evaluations
52
54
 
53
- Seven fixed tasks compare four setups using synthetic skills and correction sessions. Scores measure the share of applicable preferences followed in completed sessions, with equal weight per task. Each setup ran three repetitions; one timed-out Luna session is excluded from this view.
55
+ We tested 24 fixed engineering tasks across eight preference families, with three repetitions per setup and model. Scores measure preference adherence, with each family weighted equally.
54
56
 
55
- | Preferences followed | GPT 6.1 Sol, medium | GPT 6 Luna, high | Sonnet 5.5, high | Opus 5.5, medium |
56
- | --- | ---: | ---: | ---: | ---: |
57
- | Bare | 96.2% | 97.1% | 59.4% | 52.1% |
58
- | Existing user skills | 100% | 100% | 65.0% | 63.8% |
59
- | Skills plus Shadowclone routing | 100% | 100% | 65.0% | 63.8% |
60
- | Skills, routing, and deep learning | 100% | 100% | 81.7% | 96.4% |
57
+ | Setup | GPT 6.1 Sol | GPT 6 Luna | Claude Sonnet 5.5 | Claude Opus 5.5 |
58
+ | ----------------------------------------- | ----------: | ---------: | ----------------: | --------------: |
59
+ | Agent alone | 77.3% | 59.0% | 56.5% | 53.2% |
60
+ | Existing user skills | 85.2% | 80.6% | 76.4% | 73.6% |
61
+ | Existing skills + Shadowclone routing | 86.1% | 72.7% | 76.9% | 75.9% |
62
+ | Existing skills + handwritten preferences | 92.4% | 81.2% | 91.7% | 86.1% |
63
+ | Existing skills + Shadowclone learning | 89.4% | 83.3% | 90.0% | 82.9% |
61
64
 
62
- Learning added 16.7 percentage points over existing skills on Sonnet and 32.6 on Opus; both 95% bootstrap intervals were above zero. Both Codex models were already at 100% with skills, and routing alone tied skills. All completed tasks passed correctness and safety checks. Luna's existing-skills score covers twenty completed sessions; the other setups cover twenty-one. Its original full-matrix report remains incomplete.
65
+ Shadowclone learning improved recorded adherence over existing skills on all four models. The gain ranges from 2.8 to 13.7 percentage points. Handwritten preferences provide a reference for what agents can follow when the intended rules are supplied directly.
63
66
 
64
- The existing skill contains three preferences; one shared learning run added two from synthetic corrections. These public development tasks show a specific learning benefit, with no claim about other users, repositories, or engineering throughput. [Read the methods, intervals, timeout exclusion, and historical study](evals.md). The [eval guide](docs/guides/fixed-evals.md#how-this-follows-the-claudedev-recommendations) explains which Claude.dev recommendations the benchmark follows and the work still needed for held-out hillclimbing.
67
+ The learning setup includes routing and actual learned preferences from synthetic correction sessions. These results measure preference delivery on this task set. [Read the setup definitions, family scores, methods, and measurement limits](evals.md).
65
68
 
66
69
  ## Privacy comes first
67
70
 
@@ -69,12 +72,12 @@ The existing skill contains three preferences; one shared learning run added two
69
72
 
70
73
  **You choose what it reads.** Every capture source is off by default and has its own consent setting. Reading sessions does not silently enable memory, repository metadata, or skill editing.
71
74
 
72
- | Operation | What the selected provider can receive |
73
- | --- | --- |
74
- | Learning and skill maintenance | Selected redacted instructions, steering, and supporting context |
75
- | Browser editor | Nothing merely from opening it; optional AI drafting sends reviewed form fields |
76
- | Delegated `run` | The authorized task worktree and guidance |
77
- | Evaluation | Synthetic prompts, the tested setup's guidance, and generated code |
75
+ | Operation | What the selected provider can receive |
76
+ | ------------------------------ | ------------------------------------------------------------------------------- |
77
+ | Learning and skill maintenance | Selected redacted instructions, steering, and supporting context |
78
+ | Personal GitHub clone | The reviewed skills, native rules, selected repository, and issue or PR task |
79
+ | Browser editor | Nothing merely from opening it; optional AI drafting sends reviewed form fields |
80
+ | Evaluation | Synthetic prompts, the tested setup's guidance, and generated code |
78
81
 
79
82
  **Redaction is a boundary, not an anonymity guarantee.** Tool results, tool-returned file contents, thinking blocks, and data-access results are excluded from learning. Use only sources and repositories you are authorized to send to the selected provider.
80
83
 
@@ -82,14 +85,14 @@ The existing skill contains three preferences; one shared learning run added two
82
85
 
83
86
  ## Guides
84
87
 
85
- | Guide | Use it to |
86
- | --- | --- |
87
- | [Agent builds](docs/guides/agent-builds.md) | Choose, create, and equip skills |
88
- | [Learning](docs/guides/learning.md) | Run deep learning, background learning, or record a preference |
89
- | [Skill maintenance](docs/guides/skills.md) | Review updates, conflicts, history, and undo |
90
- | [Repository setup](docs/guides/repositories.md) | Share checks and skills with a repository |
91
- | [How it works](docs/guides/how-it-works.md) | Follow guidance from consented evidence to coding agents |
92
- | [Command reference](docs/guides/commands.md) | Find everyday and advanced commands |
88
+ | Guide | Use it to |
89
+ | ----------------------------------------------- | -------------------------------------------------------------- |
90
+ | [Agent builds](docs/guides/agent-builds.md) | Choose, create, and equip skills |
91
+ | [Learning](docs/guides/learning.md) | Run deep learning, background learning, or record a preference |
92
+ | [Skill maintenance](docs/guides/skills.md) | Review updates, conflicts, history, and undo |
93
+ | [Repository setup](docs/guides/repositories.md) | Share checks and skills with a repository |
94
+ | [How it works](docs/guides/how-it-works.md) | Follow guidance from consented evidence to coding agents |
95
+ | [Command reference](docs/guides/commands.md) | Find everyday and advanced commands |
93
96
 
94
97
  **Further reading:** [Documentation index](docs/README.md) · [Architecture](docs/architecture/README.md) · [Design history](docs/design/README.md) · [Migration](docs/migration.md) · [Contributing](CONTRIBUTING.md) · [Security](SECURITY.md)
95
98
 
package/dist/_.._/web/client/index.html CHANGED
@@ -6,7 +6,7 @@
6
6
  <meta name="referrer" content="no-referrer" />
7
7
  <meta
8
8
  http-equiv="Content-Security-Policy"
9
- content="default-src 'self'; connect-src 'self'; script-src 'self'; style-src 'self'; img-src 'self' data:; object-src 'none'; base-uri 'none'; form-action 'none'"
9
+ content="default-src 'self'; connect-src 'self'; script-src 'self'; style-src 'self'; img-src 'self' data:; object-src 'none'; base-uri 'none'; form-action https://github.com"
10
10
  />
11
11
  <title>Your agent build · Shadowclone</title>
12
12
  <link rel="icon" type="image/svg+xml" href="../../../web/shadowclone-mark-w6a76wqy.svg" />
@@ -22,7 +22,8 @@
22
22
 
23
23
 
24
24
 
25
- <link rel="stylesheet" crossorigin href="../../../web/index-dzztba9n.css"><script type="module" crossorigin src="../../../web/index-s1e5043w.js"></script></head>
25
+
26
+ <link rel="stylesheet" crossorigin href="../../../web/index-ntyv9x8a.css"><script type="module" crossorigin src="../../../web/index-twq4tm3n.js"></script></head>
26
27
  <body>
27
28
  <div class="frame">
28
29
  <header class="masthead">
@@ -38,8 +39,7 @@
38
39
  <p class="eyebrow">YOUR AGENT. YOUR INSTINCTS.</p>
39
40
  <h1 id="page-title">Build the way you think.</h1>
40
41
  <p class="subtitle">
41
- Equip a way of working. Every skill shapes how your agent approaches
42
- a task.
42
+ Equip a way of working. Every skill shapes how your agent approaches a task.
43
43
  </p>
44
44
  </section>
45
45
 
@@ -63,27 +63,28 @@
63
63
  <p id="scope-note" class="scope-note"></p>
64
64
  <details id="library-note" class="library-note" hidden>
65
65
  <summary>Built-in library · add your own</summary>
66
- <p>
67
- Include your installed skills by choosing the folders Shadowclone
68
- can read.
69
- </p>
66
+ <p>Include your installed skills by choosing the folders Shadowclone can read.</p>
70
67
  <code>shadowclone skills configure</code>
71
68
  </details>
72
69
  </div>
73
70
  <div id="notice" role="status" class="notice" hidden></div>
74
71
  <details class="requirements">
75
72
  <summary>Learning model</summary>
76
- <p>Models and credentials are configured in your coding harness. Session learning follows that session's model.</p>
73
+ <p>
74
+ Models and credentials are configured in your coding harness. Session learning follows
75
+ that session's model.
76
+ </p>
77
77
  <button id="load-models" class="quiet-button">Refresh available models</button>
78
- <label>Saved model <select id="learning-model"><option>Refresh models to choose</option></select></label>
78
+ <label
79
+ >Saved model
80
+ <select id="learning-model">
81
+ <option>Refresh models to choose</option>
82
+ </select></label
83
+ >
79
84
  </details>
80
85
 
81
86
  <div class="workspace">
82
- <section
83
- id="constellation"
84
- class="constellation"
85
- aria-label="Skill constellation"
86
- >
87
+ <section id="constellation" class="constellation" aria-label="Skill constellation">
87
88
  <div id="constellation-status" class="constellation-status" role="status">
88
89
  <p id="constellation-status-text">Loading your skills…</p>
89
90
  <button id="retry-build" class="quiet-button" hidden>Retry loading skills</button>
@@ -100,18 +101,17 @@
100
101
  Map / list
101
102
  </button>
102
103
  </div>
103
- <svg
104
- id="starfield"
105
- class="starfield"
106
- aria-hidden="true"
107
- ></svg>
104
+ <svg id="starfield" class="starfield" aria-hidden="true"></svg>
108
105
  <svg
109
106
  id="constellation-map"
110
107
  class="constellation-map"
111
108
  aria-label="Skills grouped into navigable hubs"
112
109
  ></svg>
113
110
  <div id="constellation-list" class="constellation-list"></div>
114
- <p class="tree-hint">Select a category to expand it or a skill to inspect it. Drag to pan and scroll to zoom.</p>
111
+ <p class="tree-hint">
112
+ Select a category to expand it or a skill to inspect it. Drag to pan and scroll to
113
+ zoom.
114
+ </p>
115
115
  </section>
116
116
 
117
117
  <aside class="build-panel" aria-label="Your character build">
@@ -124,20 +124,10 @@
124
124
  </p>
125
125
  </div>
126
126
  <div id="traits" class="traits"></div>
127
- <button id="describe" class="quiet-button">
128
- Describe my agent with my model
129
- </button>
130
- <div
131
- id="model-description"
132
- class="requirements"
133
- aria-live="polite"
134
- ></div>
127
+ <button id="describe" class="quiet-button">Describe my agent with my model</button>
128
+ <div id="model-description" class="requirements" aria-live="polite"></div>
135
129
  </section>
136
- <section
137
- id="detail"
138
- class="detail"
139
- aria-label="Selected skill"
140
- ></section>
130
+ <section id="detail" class="detail" aria-label="Selected skill"></section>
141
131
  </aside>
142
132
  </div>
143
133
 
@@ -147,12 +137,6 @@
147
137
  hidden
148
138
  aria-label="Repository requirements"
149
139
  ></section>
150
- <details class="requirements">
151
- <summary>Work in this repository</summary>
152
- <p>Last recorded progress. Inspect a task for current verification and review readiness.</p>
153
- <button id="refresh-tasks" class="quiet-button">Refresh recorded work</button>
154
- <div id="task-list" aria-live="polite"></div>
155
- </details>
156
140
  </main>
157
141
 
158
142
  <footer class="footer">