chati-dev 4.5.12 → 4.5.13

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.
@@ -1,8 +1,8 @@
1
1
  # chati.dev Configuration
2
- version: "4.5.12"
2
+ version: "4.5.13"
3
3
  installed_at: "2026-02-07T10:00:00Z"
4
4
  updated_at: "2026-08-27T00:00:00Z"
5
- installer_version: "4.5.12"
5
+ installer_version: "4.5.13"
6
6
  project_type: greenfield
7
7
  language: en
8
8
  ides: []
@@ -890,11 +890,11 @@ The Tasks agent assigns the kind. When a task could be either `script` or `agent
890
890
 
891
891
  ---
892
892
 
893
- *Chati.dev Constitution v4.5.12: 25 Articles + Preamble*
893
+ *Chati.dev Constitution v4.5.13: 25 Articles + Preamble*
894
894
  *v4.2.0 Amendments: Article V amended (Team Communication); Article VIII amended (Team Handoff Envelope); Article XII amended (PRISM L6 Team Roster); Article XXI added (Agent Teams Governance); Article XXII added (Fault Vector Protocol)*
895
895
  *v4.2.1 Fixes: installer refactor (recursive copy), features block in config, context files sync*
896
896
  *v4.2.2 Fixes: 10 hooks rewritten to canonical hookSpecificOutput schema; .claude/settings.json now written by installer; license enforcement 3-provider; Agent Teams default ON*
897
897
  *v4.2.3 Fixes: directory restructure (.chati.dev/ hidden + artifacts/ at root); deterministic pipeline router; Agent Teams communication validated; 3-layer integrity checking; CLI dispatch loop enforcement*
898
898
  *v4.3.0 Additions: qa-visual agent registered across all providers, model maps, artifact scopes, isolation, and entity registry; Premium Animation Standards + Visual QA Phase 5e; Article XXIII added (Reasoning Tier Governance)*
899
- *v4.5.12 Runtime: Planning seals a provider-neutral handoff; RAIL is the sole implementation and independent-review authority; Release Lane requires evidence and explicit human authorization.*
899
+ *v4.5.13 Runtime: Planning seals a provider-neutral handoff; RAIL is the sole implementation and independent-review authority; Release Lane requires evidence and explicit human authorization.*
900
900
  *All agents are bound by this Constitution. Violations are enforced per article.*
@@ -1,7 +1,7 @@
1
1
  # Chati.dev System Context
2
2
 
3
3
  ## Framework
4
- - **Version**: 4.5.12
4
+ - **Version**: 4.5.13
5
5
  - **Agents**: Specialized discovery and planning agents feeding provider-neutral RAIL execution and a sealed release lane
6
6
  - **Constitution**: 25 Articles + Preamble
7
7
  - **Quality**: 5 pipeline gates + 3-tier verdicts + Fault Vector Protocol (Article XXII)
@@ -3,7 +3,7 @@
3
3
  # and by the Health Check for system integrity validation.
4
4
 
5
5
  metadata:
6
- version: "4.5.12"
6
+ version: "4.5.13"
7
7
  last_updated: "2026-08-27T00:00:00Z"
8
8
  entity_count: 109
9
9
  checksum_algorithm: sha256
@@ -1,5 +1,5 @@
1
1
  {
2
- "createdAt": "2026-08-27T15:44:26.707Z",
2
+ "createdAt": "2026-08-27T20:02:13.399Z",
3
3
  "files": {
4
4
  "agents/build/dev.md": {
5
5
  "hash": "13c2b3855c1f815154a2f4020b99fd8ee71e701b41783f5f26696222720a1466",
@@ -97,12 +97,12 @@
97
97
  "type": "md"
98
98
  },
99
99
  "config.yaml": {
100
- "hash": "35917976892c5b3a929837f0193e6b08115c84134af233aba0c79f201688d193",
100
+ "hash": "ecdeb8f0ff8638ea702d7c6b8dfb23bf2130480fafc00767882e1cdbb97c52eb",
101
101
  "size": 2671,
102
102
  "type": "yaml"
103
103
  },
104
104
  "constitution.md": {
105
- "hash": "5b617a789e63a198828398fb673dc8ae4d9ba28728ebcabb2eaa83dc21ab1402",
105
+ "hash": "23ca8da91b0886332fcf0f6b4965fadcb0b8a6e9a0da7f5e0346d16c27387fa9",
106
106
  "size": 66739,
107
107
  "type": "md"
108
108
  },
@@ -122,12 +122,12 @@
122
122
  "type": "md"
123
123
  },
124
124
  "context/root.md": {
125
- "hash": "97f4f47c0ecb2499482401ee859512b30d9c39b53b85261c046145d4b8eeb0ae",
125
+ "hash": "1a3036f7c7831cea869b4273da77d4b4c3ddcbfcb9c0eea5fac6dd78b07a6016",
126
126
  "size": 1218,
127
127
  "type": "md"
128
128
  },
129
129
  "data/entity-registry.yaml": {
130
- "hash": "61a4cf5f7ad94b925182040befa609dc0ed8e0598f22118891337f225081667b",
130
+ "hash": "9a9ad8f246a5c751bb29523d3a24d98e7b58ac9535d75e3800538c1395ae90f9",
131
131
  "size": 34606,
132
132
  "type": "yaml"
133
133
  },
@@ -1352,5 +1352,5 @@
1352
1352
  "type": "yaml"
1353
1353
  }
1354
1354
  },
1355
- "version": "4.5.12"
1355
+ "version": "4.5.13"
1356
1356
  }
@@ -1 +1 @@
1
- qw0VeF2Ymm1U0meAoVLpTkcRUKiPWmv1LhJvOo6GdZYWc81BUGLHW1uMWcI84SkVRsI4skriATd7rAi2h13kAQ==
1
+ TEzLr1bBMHjCWgycL61RoRjockr/aGq83X862GPhOxS2eibvCeceRSM/PkUcAM4OCPAUIqmoUC/t66dhwZqhDQ==
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "chati-dev",
3
- "version": "4.5.12",
3
+ "version": "4.5.13",
4
4
  "description": "AI-Powered Multi-Agent Orchestration System - Structured vibe coding for Full Stack Development",
5
5
  "type": "module",
6
6
  "bin": {
@@ -65,6 +65,30 @@ const PROVIDER_META = {
65
65
  },
66
66
  };
67
67
 
68
+ const CODEX_CREDENTIAL_BOUNDARY = `
69
+ ## Codex external provider credential boundary
70
+
71
+ Commands returned in any \`spawn_command\`, \`parallel_spawn_command\` or
72
+ \`parallel_spawn_commands\` start another provider CLI. Codex MUST execute each
73
+ of these commands with native escalated execution so the child CLI can access
74
+ its existing local credentials:
75
+
76
+ - Set \`sandbox_permissions\` to \`require_escalated\`.
77
+ - Explain that Chati needs to run the selected provider CLI using its existing
78
+ local authentication.
79
+ - Offer the narrow reusable prefix made from the first two command arguments:
80
+ \`["node", "<absolute path from the command to run-agent.js>"]\`.
81
+ - Keep the project working directory and the router-provided command unchanged.
82
+
83
+ Never test a routed provider command inside the workspace sandbox first. A
84
+ \`PROVIDER_NOT_READY\` or \`PROVIDER_NOT_AUTHENTICATED\` result obtained inside
85
+ that sandbox is inconclusive because the sandbox can hide OS credentials. If a
86
+ sandboxed attempt already failed, rerun it once with native escalated execution.
87
+ Only a failure from that escalated attempt may be reported as an authentication
88
+ failure. Do not repeat the same sandboxed command and do not offer to skip a
89
+ required v2 pipeline stage.
90
+ `;
91
+
68
92
  // ---------------------------------------------------------------------------
69
93
  // Replacement Configuration Builder
70
94
  // ---------------------------------------------------------------------------
@@ -225,5 +249,12 @@ export function adaptFrameworkFile(content, filePath, provider) {
225
249
  result = result.replace(pattern, replacement);
226
250
  }
227
251
 
252
+ if (provider === 'codex' && filePath === 'orchestrator/chati.md') {
253
+ const routedAction = '## Action: Routed Interactive Agent';
254
+ result = result.includes(routedAction)
255
+ ? result.replace(routedAction, `${CODEX_CREDENTIAL_BOUNDARY}\n${routedAction}`)
256
+ : `${result.trimEnd()}\n\n${CODEX_CREDENTIAL_BOUNDARY.trimStart()}`;
257
+ }
258
+
228
259
  return result;
229
260
  }
@@ -229,12 +229,42 @@ name: chati
229
229
  description: >
230
230
  Use this skill when the user wants to start or continue a Chati.dev session.
231
231
  This activates the orchestrator which routes to the correct agent in the pipeline.
232
- Invoke when the user mentions chati, /chati, wants to plan, build, or deploy a project.
232
+ Invoke when the user mentions chati, /chati, or wants to discover, plan, execute, or release a project.
233
233
  Do NOT use for general coding questions unrelated to the Chati.dev pipeline.
234
234
  ---
235
235
 
236
236
  # Chati.dev Orchestrator
237
237
 
238
+ ## STEP 0 — Run the deterministic router (MUST be the first action)
239
+
240
+ Use the shell in the current project directory and run:
241
+
242
+ \`\`\`bash
243
+ node .chati.dev/orchestrator/chati-router.js
244
+ \`\`\`
245
+
246
+ Do not infer provider authentication, session state, model availability, or the
247
+ next pipeline action from files or prior messages. The router output is the
248
+ source of truth. Do not tell the user to log in unless the router or the exact
249
+ routed spawn command returns a provider authentication error.
250
+
251
+ Parse the JSON output before responding:
252
+
253
+ - If \`ok\` is \`false\`, report the exact \`error\` and relevant detail, then stop.
254
+ - If \`ok\` is \`true\`, use \`session.language\` for every response.
255
+ - Follow the returned \`action\` in the adapted orchestrator.
256
+ - When the action includes \`pipeline.spawn_command\`, execute that exact command.
257
+ Never replace it with an inferred provider check or a manually constructed command.
258
+ - A router-provided spawn command starts an external provider CLI. Execute it
259
+ with native escalated execution: set \`sandbox_permissions\` to
260
+ \`require_escalated\`, preserve the exact command and working directory, and
261
+ offer the narrow reusable prefix
262
+ \`["node", "<absolute path from spawn_command to run-agent.js>"]\`.
263
+ - Do not run that provider command inside the workspace sandbox first. A
264
+ \`PROVIDER_NOT_READY\` or \`PROVIDER_NOT_AUTHENTICATED\` result from a sandboxed
265
+ attempt is not proof that the user is logged out. Rerun it once with native
266
+ escalation before reporting an authentication failure.
267
+
238
268
  ## CRITICAL: Language Override
239
269
 
240
270
  Read \`.chati/session.yaml\` field \`language\` BEFORE anything else.