cc10x

Name: cc10x
Author: romiluz13

155stars26forks

How to install cc10x?

romiluz13/cc10x

$git clone https://github.com/romiluz13/cc10x

Installs into the current project.

Summary

The Loop Engine for Claude Code — engineer the loop, not the prompt. 1 router · 9 agents · 16 skills · 4 workflows. Fail-closed gates, test honesty, anti-anchored review.

Files · 1

View on GitHub

README.md

1	<p align="center">
2	<img src="assets/logo.png" alt="CC10x — The Loop Engine for Claude Code. Orange stylized X with starburst and CC10x wordmark." width="280" />
3	</p>
4
5	<h1 align="center">cc10x — The Loop Engine for Claude Code</h1>
6
7	<p align="center">
8	<em>Stop chasing better models. Engineer the loop.</em>
9	</p>
10
11	<p align="center">
12	Same model. Same Claude Code. Different outcome.<br />
13	The agent forgets between turns. <strong>The harness remembers across runs.</strong><br />
14	cc10x writes every workflow to disk — intent, evidence, verdicts, failures — so resume, review, and verification read from the artifact, not from a context window that's already gone.
15	</p>
16
17	<p align="center">
18	<strong>1 router</strong>  ·  <strong>11 specialist agents</strong>  ·  <strong>19 skills</strong>  ·  <strong>4 workflows</strong>
19	</p>
20
21	<p align="center">
22	Fail-closed gates  ·  survives compaction  ·  dispatch-by-reference  ·  test honesty gates  ·  anti-anchored review
23	</p>
24
25	Current version: 12.7.0
26
27	---
28
29	## Install
30
31	Step 1 — Add the marketplace:
32
33	```bash
34	/plugin marketplace add romiluz13/cc10x
35	```
36
37	Step 2 — Install the plugin:
38
39	```bash
40	/plugin install cc10x@cc10x
41	```
42
43	Then say "set up cc10x for me" in Claude Code and restart. Done.
44
45	---
46
47	## Why cc10x
48
49	Ask Claude for something complex. It works for a while. Then it declares "Done!" — tests still red, refactor half-finished, and by message 40 it's contradicting itself because the context is gone.
50
51	cc10x fixes the loop, not the prompt. A better model running free loses to the same model, constrained and looped correctly. That's the whole bet.
52
53	\| The pain you know \| How cc10x handles it \|
54	\| --- \| --- \|
55	\| "Done!" on red tests \| `integration-verifier` is independent of the builder. Phase-exit gates block advancement on partial evidence. \|
56	\| Silent failures nobody asked about \| `failure-hunter` runs in parallel with review — greps for swallowed errors and empty catches. \|
57	\| Context falls apart after compaction \| Workflow state on disk with stable UUIDs. Memory files the router auto-heals. \|
58	\| Planning is just a chat \| Three planning modes chosen by intent, with a fresh anti-anchored review by a reviewer who never saw the planner's rationale. \|
59	\| 12 slash commands to remember \| One router. Every request hits `cc10x-router` first. \|
60	\| `.claude/` prompt spam on every fanout \| State lives at `.cc10x/` — outside `.claude/`, so the harness's sensitive-file gate never fires. \|
61	\| Green tests that prove nothing \| Test Honesty Gates grep for `getByTestId('…-mock')`, `as any`, `.find()` bypass, `setTimeout()` waits. A hit can't count as PASS on that test's strength alone. \|
62	\| Reviewer findings that sound right but aren't \| Every finding at confidence ≥80 needs a verbatim `file:line` quote or it's auto-demoted. The verifier independently re-reads the line and drops hallucinated findings before they can gate. \|
63	\| Orchestrator context rotting with pasted history \| Dispatch by reference, not by blob — the diff is written to disk, the prompt passes a path, never a body. (One real dispatch hit 42k chars, 99% pasted history. The scar became law.) \|
64
65	Everything below is the architecture. Keep reading if you want to know how before you install.
66
67	---
68
69	## Runtime Model
70
71	### 1. Router owns orchestration
72
73	`cc10x-router` is the only orchestration authority.
74
75	The router now uses a kernel + mandatory reference shape:
76
77	- universal orchestration law stays inline in `cc10x-router/SKILL.md`
78	- workflow-specific playbooks and appendix-heavy artifact/remediation law live in `cc10x-router/references/*.md`
79	- the kernel explicitly tells Claude which reference must be read before BUILD / DEBUG / REVIEW / PLAN branch logic continues
80
81	That keeps orchestration salient without turning the router into a context dump.
82
83	It decides:
84
85	- which workflow to run
86	- which subagent to invoke next
87	- when to pause for clarification or scope decisions
88	- when remediation is required
89	- when a workflow is allowed to advance
90	- when memory and workflow artifacts are finalized
91
92	Agents do not own workflow state. They return structured results. The router interprets them.
93
94	### 2. Agents are narrow specialists
95
96	The shipped subagents are intentionally specialized:
97
98	- `planner`
99	- `plan-gap-reviewer`
100	- `component-builder`
101	- `bug-investigator`
102	- `code-reviewer`
103	- `failure-hunter`
104	- `integration-verifier`
105	- `researcher`
106	- `doc-syncer`
107
108	Each agent is optimized for one role. This keeps prompts sharper and makes workflow behavior easier to reason about.
109
110	### 3. Skills are reusable local instructions
111
112	Skills are the reusable instruction layer that agents and the router depend on.
113
114	They provide:
115
116	- planning patterns
117	- TDD rules
118	- debugging patterns
119	- review rules
120	- research synthesis
121	- memory handling
122	- verification-before-completion discipline
123
124	### 4. Workflow artifacts are the durable truth
125
126	cc10x writes proof-of-work workflow state under:
127
128	```text

People who liked this also viewed

6 picks

Type

claude-codeClaude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.WorkflowsJul 2026—139k
TrellisThe best agent harness.WorkflowsJul 2026—13k
reflyThe first open-source agent skills builder. Define skills by vibe workflow, run on Claude Code, Cursor, Codex & more. Build Clawdbot 🦞· APIs for Lovable · Bots for Slack & Lark/Feishu · Skills are infrastructure, not prompts.WorkflowsMar 2026—7.5k
pyspurA visual playground for agentic workflows: Iterate over your agents 10x fasterWorkflowsJun 2026—5.8k
claude-code🚀 Open source Claude Code CLI source code. Advanced AI Agent for developers. Includes TypeScript codebase for LLM tool-calling, agentic workflows, and terminal UI. Remember this is just the skeleton not the brain itself. Found by Chaofan Shou. WorkflowsApr 2026—3.8k
claude-codeClaude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.WorkflowsApr 2026—3.1k

cc10x

155stars26forks

View on GitHub

How to install cc10x?

romiluz13/cc10x

$git clone https://github.com/romiluz13/cc10x

Installs into the current project.

Summary

The Loop Engine for Claude Code — engineer the loop, not the prompt. 1 router · 9 agents · 16 skills · 4 workflows. Fail-closed gates, test honesty, anti-anchored review.

Files · 1

View on GitHub

README.md

1	<p align="center">
2	<img src="assets/logo.png" alt="CC10x — The Loop Engine for Claude Code. Orange stylized X with starburst and CC10x wordmark." width="280" />
3	</p>
4
5	<h1 align="center">cc10x — The Loop Engine for Claude Code</h1>
6
7	<p align="center">
8	<em>Stop chasing better models. Engineer the loop.</em>
9	</p>
10
11	<p align="center">
12	Same model. Same Claude Code. Different outcome.<br />
13	The agent forgets between turns. <strong>The harness remembers across runs.</strong><br />
14	cc10x writes every workflow to disk — intent, evidence, verdicts, failures — so resume, review, and verification read from the artifact, not from a context window that's already gone.
15	</p>
16
17	<p align="center">
18	<strong>1 router</strong>  ·  <strong>11 specialist agents</strong>  ·  <strong>19 skills</strong>  ·  <strong>4 workflows</strong>
19	</p>
20
21	<p align="center">
22	Fail-closed gates  ·  survives compaction  ·  dispatch-by-reference  ·  test honesty gates  ·  anti-anchored review
23	</p>
24
25	Current version: 12.7.0
26
27	---
28
29	## Install
30
31	Step 1 — Add the marketplace:
32
33	```bash
34	/plugin marketplace add romiluz13/cc10x
35	```
36
37	Step 2 — Install the plugin:
38
39	```bash
40	/plugin install cc10x@cc10x
41	```
42
43	Then say "set up cc10x for me" in Claude Code and restart. Done.
44
45	---
46
47	## Why cc10x
48
49	Ask Claude for something complex. It works for a while. Then it declares "Done!" — tests still red, refactor half-finished, and by message 40 it's contradicting itself because the context is gone.
50
51	cc10x fixes the loop, not the prompt. A better model running free loses to the same model, constrained and looped correctly. That's the whole bet.
52
53	\| The pain you know \| How cc10x handles it \|
54	\| --- \| --- \|
55	\| "Done!" on red tests \| `integration-verifier` is independent of the builder. Phase-exit gates block advancement on partial evidence. \|
56	\| Silent failures nobody asked about \| `failure-hunter` runs in parallel with review — greps for swallowed errors and empty catches. \|
57	\| Context falls apart after compaction \| Workflow state on disk with stable UUIDs. Memory files the router auto-heals. \|
58	\| Planning is just a chat \| Three planning modes chosen by intent, with a fresh anti-anchored review by a reviewer who never saw the planner's rationale. \|
59	\| 12 slash commands to remember \| One router. Every request hits `cc10x-router` first. \|
60	\| `.claude/` prompt spam on every fanout \| State lives at `.cc10x/` — outside `.claude/`, so the harness's sensitive-file gate never fires. \|
61	\| Green tests that prove nothing \| Test Honesty Gates grep for `getByTestId('…-mock')`, `as any`, `.find()` bypass, `setTimeout()` waits. A hit can't count as PASS on that test's strength alone. \|
62	\| Reviewer findings that sound right but aren't \| Every finding at confidence ≥80 needs a verbatim `file:line` quote or it's auto-demoted. The verifier independently re-reads the line and drops hallucinated findings before they can gate. \|
63	\| Orchestrator context rotting with pasted history \| Dispatch by reference, not by blob — the diff is written to disk, the prompt passes a path, never a body. (One real dispatch hit 42k chars, 99% pasted history. The scar became law.) \|
64
65	Everything below is the architecture. Keep reading if you want to know how before you install.
66
67	---
68
69	## Runtime Model
70
71	### 1. Router owns orchestration
72
73	`cc10x-router` is the only orchestration authority.
74
75	The router now uses a kernel + mandatory reference shape:
76
77	- universal orchestration law stays inline in `cc10x-router/SKILL.md`
78	- workflow-specific playbooks and appendix-heavy artifact/remediation law live in `cc10x-router/references/*.md`
79	- the kernel explicitly tells Claude which reference must be read before BUILD / DEBUG / REVIEW / PLAN branch logic continues
80
81	That keeps orchestration salient without turning the router into a context dump.
82
83	It decides:
84
85	- which workflow to run
86	- which subagent to invoke next
87	- when to pause for clarification or scope decisions
88	- when remediation is required
89	- when a workflow is allowed to advance
90	- when memory and workflow artifacts are finalized
91
92	Agents do not own workflow state. They return structured results. The router interprets them.
93
94	### 2. Agents are narrow specialists
95
96	The shipped subagents are intentionally specialized:
97
98	- `planner`
99	- `plan-gap-reviewer`
100	- `component-builder`
101	- `bug-investigator`
102	- `code-reviewer`
103	- `failure-hunter`
104	- `integration-verifier`
105	- `researcher`
106	- `doc-syncer`
107
108	Each agent is optimized for one role. This keeps prompts sharper and makes workflow behavior easier to reason about.
109
110	### 3. Skills are reusable local instructions
111
112	Skills are the reusable instruction layer that agents and the router depend on.
113
114	They provide:
115
116	- planning patterns
117	- TDD rules
118	- debugging patterns
119	- review rules
120	- research synthesis
121	- memory handling
122	- verification-before-completion discipline
123
124	### 4. Workflow artifacts are the durable truth
125
126	cc10x writes proof-of-work workflow state under:
127
128	```text

People who liked this also viewed

6 picks

Type