pantheon-opencode 2.1.0 → 2.1.2
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/AGENTS.global.md +2 -2
- package/AGENTS.md +1 -1
- package/CHANGELOG.md +40 -0
- package/README.md +19 -6
- package/README.pt-BR.md +19 -6
- package/package.json +1 -1
- package/plugin.json +1 -1
- package/pyproject.toml +1 -1
- package/scripts/docs-media.mjs +1 -1
- package/skills-lock.json +1 -1
- package/src/agents/apollo.md +3 -1
- package/src/agents/athena.md +11 -10
- package/src/agents/themis.md +1 -1
- package/src/agents/zeus.md +6 -4
- package/src/instructions/agent-return-format.instructions.md +1 -1
- package/src/pantheon/native-vision.ts +1 -1
- package/src/plugin/index.js +3 -3
- package/src/plugin.ts +2 -2
- package/src/plugins/tui/dist/tui.js +5 -5
- package/src/plugins/tui/package-lock.json +2 -2
- package/src/plugins/tui/package.json +1 -1
- package/src/skills/prompt-refinery/SKILL.md +6 -4
package/AGENTS.global.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
<!-- pantheon:generated:begin — Pantheon-managed region; regenerated on install/upgrade, never edit inside -->
|
|
2
2
|
<!-- Generated by scripts/build-agents-md.mjs — do not edit directly. -->
|
|
3
3
|
<!-- Source of truth: src/instructions/*.instructions.md + the static sections in this script. -->
|
|
4
|
-
<!-- pantheon:agents-md v2 artifact=global generated-at=2026-11-
|
|
4
|
+
<!-- pantheon:agents-md v2 artifact=global generated-at=2026-11-13T10:55:59-03:00 sha=7d3bb91 -->
|
|
5
5
|
|
|
6
6
|
# Pantheon Agent System — OpenCode
|
|
7
7
|
|
|
@@ -48,7 +48,7 @@ For investigation agents (Apollo, Athena), return structured findings: key findi
|
|
|
48
48
|
|
|
49
49
|
### Athena Plan Handoff
|
|
50
50
|
|
|
51
|
-
When Zeus delegates Athena through the root `subagent` dispatch, its final task result MUST contain the complete Markdown plan (not a summary, artifact path, or approval question): goal, scope and assumptions, phases, affected files or areas, dependencies, verification, risks, open questions, acceptance criteria — unknowns marked, never guessed. This supersedes the structured-findings format. Athena must not call `agent/askQuestions`, pause for user input, or ask for approval in the child session. Zeus publishes the full plan
|
|
51
|
+
When Zeus delegates Athena through the root `subagent` dispatch, its final task result MUST contain the complete Markdown plan (not a summary, artifact path, or approval question): goal, scope and assumptions, phases, affected files or areas, dependencies, verification, risks, open questions, acceptance criteria — unknowns marked, never guessed. This supersedes the structured-findings format. Athena must not call `agent/askQuestions`, pause for user input, or ask for approval in the child session. Zeus publishes the full plan before approval UI, never bundling them; if order is not guaranteed, Zeus asks in plain text afterwards. For Open Questions or decisions needed, Zeus prompts via the native `question` tool with structured options and marked recommended defaults, never leaving inert text bullets. Direct Athena invocations show the complete plan before requesting approval.
|
|
52
52
|
|
|
53
53
|
## Council Specialist Response Format
|
|
54
54
|
|
package/AGENTS.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
<!-- pantheon:generated:begin — Pantheon-managed region; regenerated on install/upgrade, never edit inside -->
|
|
2
2
|
<!-- Generated by scripts/build-agents-md.mjs — do not edit directly. -->
|
|
3
3
|
<!-- Source of truth: project sections in scripts/build-agents-md.mjs. -->
|
|
4
|
-
<!-- pantheon:agents-md v2 artifact=project generated-at=2026-11-
|
|
4
|
+
<!-- pantheon:agents-md v2 artifact=project generated-at=2026-11-13T10:55:59-03:00 sha=7d3bb91 -->
|
|
5
5
|
|
|
6
6
|
# Pantheon — Repository Setup
|
|
7
7
|
|
package/CHANGELOG.md
CHANGED
|
@@ -13,6 +13,46 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
|
|
|
13
13
|
`node scripts/versioning.mjs apply --notes` pre-fills the generated sections from conventional commits.
|
|
14
14
|
Running apply will move this section to a versioned entry and reset the template below. -->
|
|
15
15
|
|
|
16
|
+
## [v2.1.2] - 2026-11-14
|
|
17
|
+
|
|
18
|
+
### 🆕 What's New
|
|
19
|
+
|
|
20
|
+
- **Search & Worktree Freedom**: Removed search and read limits on .pantheon and active worktrees across Apollo, Themis, Athena, and Zeus.
|
|
21
|
+
- **Bilingual Portal & SVG Diagrams**: Added native EN/PT-BR language switcher and interactive Mermaid SVG diagram rendering to documentation.
|
|
22
|
+
- **Overhauled Quick Start**: Clarified Council vs Deepwork command roles, Fast-Track bypass, and interactive question gate in EN and PT-BR.
|
|
23
|
+
|
|
24
|
+
### Added
|
|
25
|
+
|
|
26
|
+
- **docs** — added `pt-BR/COMPATIBILITY.md`, expanded `pt-BR/INSTALLATION.md` with CLI flag matrix and flowchart, and expanded `cookbook.md` to 12 practical recipes in EN and PT-BR.
|
|
27
|
+
|
|
28
|
+
### Changed
|
|
29
|
+
|
|
30
|
+
- **docs** — revised Quick Start in EN and PT-BR with accurate routing, Fast-Track vs Deepwork distinction, native question gate flow, and Mermaid execution lifecycle diagrams.
|
|
31
|
+
|
|
32
|
+
### Fixed
|
|
33
|
+
|
|
34
|
+
- **agents** — removed all search and read restrictions on the `.pantheon/` folder and active worktrees in `@apollo`, `@themis`, `@athena`, and `@zeus`, excluding only `.git` and `node_modules`.
|
|
35
|
+
|
|
36
|
+
## [v2.1.1] - 2026-11-13
|
|
37
|
+
|
|
38
|
+
### 🆕 What's New
|
|
39
|
+
|
|
40
|
+
- **Interactive Question Gate**: Native OpenCode `question` tool strictly enforced for all open decisions in planning, with marked defaults.
|
|
41
|
+
- **Runnable Real-World Examples**: Complete production examples (`examples/saas-dashboard` & `examples/cli-tool`) with automated test suites.
|
|
42
|
+
- **Comprehensive Docs & Wiki**: Modernized guides portal (`docs/guides/`), interactive Mermaid capability matrix, and GitHub Wiki mirror.
|
|
43
|
+
- **V2 Legacy Cleanup**: Removed obsolete "V2" naming clutter across internal policy markers, visual context tags, tape scripts, and docs.
|
|
44
|
+
|
|
45
|
+
### Added
|
|
46
|
+
|
|
47
|
+
- **docs** — modern guide collection (`first-project.md`, `deepwork.md`, `vision.md`, `cookbook.md`), bilingual README tables, and interactive Mermaid decision diagrams.
|
|
48
|
+
- **examples** — `saas-dashboard` (React 19 + Fastify + CSV streaming + canonical Deepwork plan) and `cli-tool` (repo health checker CLI with tests).
|
|
49
|
+
- **wiki** — comprehensive GitHub Wiki mirror in `wiki/` (`Home.md`, `Agent-Matrix.md`, `Deepwork-Guide.md`, `Pantheon-Vision.md`, `FAQ.md`).
|
|
50
|
+
|
|
51
|
+
### Changed
|
|
52
|
+
|
|
53
|
+
- **agents** — updated `@zeus`, `@athena`, and `prompt-refinery` to forbid dumping open questions into chat; required interactive `question` tool with marked `(Recomendado)` defaults.
|
|
54
|
+
- **core** — renamed internal markers to `<!-- pantheon-policy -->` and `<!-- pantheon-visual-context -->`; cleaned up legacy V2 file names.
|
|
55
|
+
|
|
16
56
|
## [v2.1.0] - 2026-11-12
|
|
17
57
|
|
|
18
58
|
### 🆕 What's New
|
package/README.md
CHANGED
|
@@ -9,6 +9,7 @@ and teams who want useful structure without giving up speed or control of their
|
|
|
9
9
|
[Repository](https://github.com/ils15/pantheon-opencode) · [MIT License](LICENSE)
|
|
10
10
|
|
|
11
11
|
[](https://github.com/ils15/pantheon-opencode/releases/latest)
|
|
12
|
+
[](https://ils15.github.io/pantheon-opencode/)
|
|
12
13
|
[](https://github.com/ils15/pantheon-opencode/actions)
|
|
13
14
|
[](https://doi.org/10.5281/zenodo.22650136)
|
|
14
15
|
|
|
@@ -201,14 +202,26 @@ Run `/pantheon-settings` (or `Pantheon: Settings` in the command palette) to ope
|
|
|
201
202
|
|
|
202
203
|
---
|
|
203
204
|
|
|
204
|
-
## Documentation
|
|
205
|
+
## Documentation & Tutorials
|
|
205
206
|
|
|
206
|
-
|
|
207
|
+
> 📖 **Official Documentation Portal:** Explore full searchable guides, interactive Mermaid diagrams, and versioned releases at **[https://ils15.github.io/pantheon-opencode/](https://ils15.github.io/pantheon-opencode/)**.
|
|
208
|
+
|
|
209
|
+
| Guides & Tutorials | Architecture & Deep Dives | Reference & Operations |
|
|
207
210
|
|---|---|---|
|
|
208
|
-
| [
|
|
209
|
-
| [
|
|
210
|
-
| [
|
|
211
|
-
| [
|
|
211
|
+
| [First Project (End-to-End)](docs/guides/first-project.md) | [Architecture](docs/ARCHITECTURE.md) | [MCP Tools](docs/mcp-tools.md) |
|
|
212
|
+
| [Agent Capability Matrix](docs/AGENT-CAPABILITY-MATRIX.md) | [Delegation & Gates](docs/PERMISSIONS-QUALITY-DELEGATION.md) | [Native Decisions (Jev)](docs/reference/native-decisions.md) |
|
|
213
|
+
| [Deepwork Blueprint](docs/guides/deepwork.md) | [Memory System](docs/MEMORY.md) | [Model Routing](docs/reference/model-routing.md) |
|
|
214
|
+
| [Pantheon Vision](docs/guides/vision.md) | [OpenCode Integration](docs/opencode.md) | [Changelog](CHANGELOG.md) |
|
|
215
|
+
| [Cookbook](docs/guides/cookbook.md) | [Compatibility](docs/COMPATIBILITY.md) | [Wiki](wiki/Home.md) |
|
|
216
|
+
|
|
217
|
+
---
|
|
218
|
+
|
|
219
|
+
## Real-World Examples
|
|
220
|
+
|
|
221
|
+
Working runnable reference projects demonstrating Pantheon orchestration in practice:
|
|
222
|
+
|
|
223
|
+
* **[SaaS Performance Dashboard](examples/saas-dashboard/) (`examples/saas-dashboard`):** Fullstack web application with React 19, Fastify backend, streaming CSV export, automated tests, and a real Deepwork execution plan (`plan.json`, `journal.ndjson`, `PLAN.md`).
|
|
224
|
+
* **[Repo Health CLI](examples/cli-tool/) (`examples/cli-tool`):** Sub-second diagnostic CLI tool demonstrating the fast-path execution of `@talos` and `@hermes`.
|
|
212
225
|
|
|
213
226
|
---
|
|
214
227
|
|
package/README.pt-BR.md
CHANGED
|
@@ -8,6 +8,7 @@ O Pantheon reúne planejamento, implementação, revisão e documentação numa
|
|
|
8
8
|
[Repositório](https://github.com/ils15/pantheon-opencode) · [Licença MIT](LICENSE)
|
|
9
9
|
|
|
10
10
|
[](https://github.com/ils15/pantheon-opencode/releases/latest)
|
|
11
|
+
[](https://ils15.github.io/pantheon-opencode/)
|
|
11
12
|
[](https://github.com/ils15/pantheon-opencode/actions)
|
|
12
13
|
[](https://doi.org/10.5281/zenodo.22650136)
|
|
13
14
|
|
|
@@ -197,14 +198,26 @@ Execute `/pantheon-settings` (ou `Pantheon: Settings` na paleta de comandos) par
|
|
|
197
198
|
|
|
198
199
|
---
|
|
199
200
|
|
|
200
|
-
## Documentação
|
|
201
|
+
## Documentação & Tutoriais
|
|
201
202
|
|
|
202
|
-
|
|
203
|
+
> 📖 **Portal Oficial de Documentação:** Explore guias completos com pesquisa instantânea, diagramas interativos em Mermaid e histórico de versões em **[https://ils15.github.io/pantheon-opencode/](https://ils15.github.io/pantheon-opencode/)**.
|
|
204
|
+
|
|
205
|
+
| Guias & Tutoriais | Arquitetura e Aprofundamento | Referência e Operações |
|
|
203
206
|
|---|---|---|
|
|
204
|
-
| [
|
|
205
|
-
| [
|
|
206
|
-
| [
|
|
207
|
-
| [
|
|
207
|
+
| [Primeiro Projeto (End-to-End)](docs/guides/first-project.md) | [Arquitetura](docs/ARCHITECTURE.md) | [Ferramentas MCP](docs/mcp-tools.md) |
|
|
208
|
+
| [Matriz de Agentes](docs/AGENT-CAPABILITY-MATRIX.md) | [Delegação & Gates](docs/PERMISSIONS-QUALITY-DELEGATION.md) | [Decisões Nativas (Jev)](docs/reference/native-decisions.md) |
|
|
209
|
+
| [Guia do Deepwork](docs/guides/deepwork.md) | [Sistema de Memória](docs/MEMORY.md) | [Roteamento de Modelos](docs/reference/model-routing.md) |
|
|
210
|
+
| [Pantheon Vision](docs/guides/vision.md) | [Integração OpenCode](docs/opencode.md) | [Changelog](CHANGELOG.md) |
|
|
211
|
+
| [Cookbook de Receitas](docs/guides/cookbook.md) | [Compatibilidade](docs/COMPATIBILITY.md) | [Wiki](wiki/Home.md) |
|
|
212
|
+
|
|
213
|
+
---
|
|
214
|
+
|
|
215
|
+
## Exemplos Práticos do Mundo Real
|
|
216
|
+
|
|
217
|
+
Projetos funcionais demonstrando a orquestração do Pantheon na prática:
|
|
218
|
+
|
|
219
|
+
* **[SaaS Performance Dashboard](examples/saas-dashboard/) (`examples/saas-dashboard`):** Aplicação fullstack completa com React 19, backend Fastify, streaming de CSV, testes automatizados e plano Deepwork real com evidências (`plan.json`, `journal.ndjson`, `PLAN.md`).
|
|
220
|
+
* **[Repo Health CLI](examples/cli-tool/) (`examples/cli-tool`):** Ferramenta de linha de comando para diagnóstico de repositórios em menos de 1 segundo, demonstrando o fluxo fast-path do `@talos` e `@hermes`.
|
|
208
221
|
|
|
209
222
|
---
|
|
210
223
|
|
package/package.json
CHANGED
package/plugin.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "pantheon",
|
|
3
3
|
"publisher": "ils15",
|
|
4
|
-
"version": "2.1.
|
|
4
|
+
"version": "2.1.2",
|
|
5
5
|
"displayName": "Pantheon",
|
|
6
6
|
"description": "Conductor-Delegate orchestration platform with 14 specialized agents, 25 skills, and OpenCode support.",
|
|
7
7
|
"agents": "./src/agents",
|
package/pyproject.toml
CHANGED
|
@@ -4,7 +4,7 @@ build-backend = "setuptools.backends._legacy:_Backend"
|
|
|
4
4
|
|
|
5
5
|
[project]
|
|
6
6
|
name = "pantheon"
|
|
7
|
-
version = "2.1.
|
|
7
|
+
version = "2.1.2"
|
|
8
8
|
description = "Multi-agent orchestration framework"
|
|
9
9
|
requires-python = ">=3.11"
|
|
10
10
|
# Python sources under src/mcp/ import fastmcp/pydantic at runtime, but those
|
package/scripts/docs-media.mjs
CHANGED
|
@@ -9,7 +9,7 @@
|
|
|
9
9
|
*
|
|
10
10
|
* Usage:
|
|
11
11
|
* node scripts/docs-media.mjs # record every *.tape
|
|
12
|
-
* node scripts/docs-media.mjs
|
|
12
|
+
* node scripts/docs-media.mjs sidebar # record one tape (name or file)
|
|
13
13
|
* node scripts/docs-media.mjs --check # verify prerequisites only
|
|
14
14
|
*/
|
|
15
15
|
|
package/skills-lock.json
CHANGED
|
@@ -23,7 +23,7 @@
|
|
|
23
23
|
"src/skills/memory-bank/references/storage-model.md": "728a9e5f2055ce0f74c759405072a6fe57f65587bdb7b9e5e7fa90b60528b085",
|
|
24
24
|
"src/skills/memory-bank/references/structure-and-frequency.md": "09caedf905870c688401a2e58b2776a9c6187fe724c576019c58c3b80fe1b98f",
|
|
25
25
|
"src/skills/memory-bank/SKILL.md": "047b3049c27e908b1d79afb14f79336e32f775385ba23956771166d445944f2f",
|
|
26
|
-
"src/skills/prompt-refinery/SKILL.md": "
|
|
26
|
+
"src/skills/prompt-refinery/SKILL.md": "61140e0677d4d94432e79aad37dfd2d846b65fa80431d881d3759d05334157b6",
|
|
27
27
|
"src/skills/reflect/SKILL.md": "297c374f1fbc8e15d3848f51e811e0614d4a008c7d0e1daf88bbb240d605b8a8",
|
|
28
28
|
"src/skills/security-hardening/SKILL.md": "6d637e122421f7bc3b751f04854ad343a5a5d35249458bc8939233138fd3f20d",
|
|
29
29
|
"src/skills/session-goal/references/goal-format.md": "071f6f8f12cfbc26d4703cd5f34f839f91e586a3b969bf5f81a74b961c254a83",
|
package/src/agents/apollo.md
CHANGED
|
@@ -63,6 +63,8 @@ mcp_tools:
|
|
|
63
63
|
|
|
64
64
|
- **Strictly read-only scout**: search with grep/glob/read and `pantheon-memory` (`memory_search`, `code_query`). Never modify files, never run shell commands.
|
|
65
65
|
- **Search order**: session context first, direct targeted searches second. Batch independent queries only when discovery scope benefits from parallel search.
|
|
66
|
-
- **
|
|
66
|
+
- **Search Scope & Permissions**:
|
|
67
|
+
- **Permitted targets**: Targeted searches and reads across codebase files (`src/`, `tests/`, `docs/`, etc.), the `.pantheon/` folder (`.pantheon/**`, including deepwork plans, memory-bank, reports, and code-mode scripts), and active worktrees (`.pantheon-worktrees/**`) are fully permitted without restriction whenever relevant to the task or requested by the user.
|
|
68
|
+
- **Excluded noise**: Avoid broad, unbounded recursive scans across `.git/**` or `node_modules/**`.
|
|
67
69
|
- **External research**: Use configured web search MCP or `webfetch` for documentation URLs. Never invent citations.
|
|
68
70
|
- **Output format**: Return hierarchical structured findings (`files_changed`, `summary`, `confidence`). No raw dumps.
|
package/src/agents/athena.md
CHANGED
|
@@ -91,7 +91,7 @@ Use `webfetch` to verify availability before proposing model updates to other ag
|
|
|
91
91
|
|
|
92
92
|
**Rules**:
|
|
93
93
|
- Max 1-2 direct codebase searches in src/ or docs/; only a root Athena may delegate additional discovery to @apollo
|
|
94
|
-
-
|
|
94
|
+
- Inspecting codebase files, the `.pantheon/` directory (`.pantheon/**`), and active worktrees (`.pantheon-worktrees/**`) is fully permitted without restriction. Avoid broad recursive scans across `.git/**` or `node_modules/**`.
|
|
95
95
|
- Convergence rule: 80% understanding OR stop immediately (fast bounded planning)
|
|
96
96
|
- Simple features: Direct search + plan (no Apollo)
|
|
97
97
|
- Complex features: 1-2 targeted searches, then Apollo discovery only from a root Athena; a child reports remaining discovery needs to Zeus
|
|
@@ -110,7 +110,7 @@ Use `webfetch` to verify availability before proposing model updates to other ag
|
|
|
110
110
|
- Spend time re-planning or iterating beyond 5 min
|
|
111
111
|
- Wait for perfect understanding
|
|
112
112
|
- Make multiple planning attempts
|
|
113
|
-
-
|
|
113
|
+
- Perform broad, unbounded recursive scans across `.git/**` or `node_modules/**`
|
|
114
114
|
|
|
115
115
|
## Plan Structure (Didactic Executive Summary + Disk Blueprint)
|
|
116
116
|
|
|
@@ -166,7 +166,9 @@ flowchart LR
|
|
|
166
166
|
```
|
|
167
167
|
|
|
168
168
|
### Interactive Question Gate
|
|
169
|
-
When directly invoked by the user
|
|
169
|
+
When directly invoked by the user:
|
|
170
|
+
- If the plan contains any Open Questions (`decisions_needed`), **NEVER leave them as plain text bullets in chat**. You MUST call the native `question` tool, decomposing every open question into an interactive multiple-choice item with concrete options (marking recommended choice with `(Recomendado)`).
|
|
171
|
+
- Alongside open questions, always include the approval question with structured choices (e.g., `["(Recomendado) Aprovar e executar", "Ajustar plano"]`), rather than relying on unstructured chat text.
|
|
170
172
|
Keep unknown paths and unverified assumptions explicit rather than fabricating
|
|
171
173
|
detail.
|
|
172
174
|
|
|
@@ -196,14 +198,13 @@ approval.
|
|
|
196
198
|
|
|
197
199
|
## Approval Gate
|
|
198
200
|
|
|
199
|
-
After the complete plan is visible in the current conversation, use
|
|
200
|
-
|
|
201
|
-
|
|
202
|
-
|
|
203
|
-
-
|
|
204
|
-
```
|
|
201
|
+
After the complete plan is visible in the current conversation, use the native `question` tool for direct user invocations:
|
|
202
|
+
- Decompose all Open Questions into interactive questions with concrete choices and marked `(Recomendado)` defaults.
|
|
203
|
+
- Gate Question:
|
|
204
|
+
- Question: "Aprovar plano e iniciar execução?"
|
|
205
|
+
- Options: ["(Recomendado) Aprovar e executar", "Ajustar plano"]
|
|
205
206
|
|
|
206
|
-
Only after explicit
|
|
207
|
+
Only after explicit user approval → delegate to @zeus with plan context.
|
|
207
208
|
|
|
208
209
|
## When to Use Apollo
|
|
209
210
|
|
package/src/agents/themis.md
CHANGED
|
@@ -64,7 +64,7 @@ Quality & security gate for all Pantheon agents. Three-layer review with IntentG
|
|
|
64
64
|
## Fast-Track Bypass
|
|
65
65
|
- Bounded changes (≤2 files, ≤50 lines) NEVER go through Themis. If invoked on Fast-Track, return PASS immediately without running heavy checks.
|
|
66
66
|
- NEVER run the entire test suite. Run only the narrow test slice via `test-slicing`.
|
|
67
|
-
-
|
|
67
|
+
- Scanning and auditing code inside active worktrees (`.pantheon-worktrees/**`), the `.pantheon/` directory (`.pantheon/**`), and Deepwork verification criteria is fully permitted without restriction. Avoid broad recursive scans across `.git/**` or `node_modules/**`.
|
|
68
68
|
|
|
69
69
|
## IntentGate
|
|
70
70
|
|
package/src/agents/zeus.md
CHANGED
|
@@ -55,10 +55,10 @@ skills:
|
|
|
55
55
|
- In Scope, Out of Scope & Negative Constraints
|
|
56
56
|
- Acceptance Criteria
|
|
57
57
|
- Open Questions with Recommended Defaults
|
|
58
|
-
- **MANDATORY `question` tool:**
|
|
58
|
+
- **MANDATORY `question` tool:** When Open Questions, decisions needed, or scope trade-offs are present, ALWAYS invoke the native `question` tool with structured choices (marking the recommended choice with `(Recomendado)`) alongside the execution approval question (e.g. `["(Recomendado) Aprovar escopo e iniciar execução", "Ajustar escopo/premissas"]`). NEVER leave open questions as inert text bullets in chat and proceed blindly.
|
|
59
59
|
- **Ambiguous or Multi-Concern Prompts (Tier 1 Silent Lean):**
|
|
60
60
|
- When a request mixes refactor/fix/redesign without clear acceptance criteria, extract internal lean brief (Target, Slices, Constraints, Verification, Defaults).
|
|
61
|
-
- If ambiguity blocks safe execution or presents critical trade-offs, **always invoke the native `question` tool** with concrete options rather than making blind assumptions or writing plain-text questions.
|
|
61
|
+
- If ambiguity blocks safe execution or presents critical trade-offs, **always invoke the native `question` tool** with concrete options (recommended option first) rather than making blind assumptions or writing plain-text questions.
|
|
62
62
|
|
|
63
63
|
## Fast-Path Direct Implementation (Default Coding Workflow)
|
|
64
64
|
|
|
@@ -74,7 +74,7 @@ skills:
|
|
|
74
74
|
|
|
75
75
|
## Golden Rule & Boundaries
|
|
76
76
|
|
|
77
|
-
**Coordinator and router.** Zeus never edits or implements code (delegates to @hermes, @aphrodite, @talos, etc.), but may perform bounded reads (≤50 lines) for fast routing without unnecessary subagent dispatches. Broad investigations or exploratory research go to @apollo.
|
|
77
|
+
**Coordinator and router.** Zeus never edits or implements code (delegates to @hermes, @aphrodite, @talos, etc.), but may perform bounded reads (≤50 lines) for fast routing without unnecessary subagent dispatches. Broad investigations or exploratory research go to @apollo. Routing and inspecting files across the `.pantheon/` directory (`.pantheon/**`, including plans, memory-bank, reports, configs) and active worktrees (`.pantheon-worktrees/**`) is fully permitted without restriction. Avoid broad recursive scans across `.git/**` or `node_modules/**`.
|
|
78
78
|
|
|
79
79
|
Zeus NEVER edits files (`edit: deny`, `hashline_edit: deny`). Zero shell access (`bash: deny`, `shell: deny`).
|
|
80
80
|
NEVER ask the user to type terminal commands — dispatch the specialist with bash permitted (@prometheus, @hermes, @talos).
|
|
@@ -133,5 +133,7 @@ Whenever the user explicitly asks for a plan, or starts `/pantheon-deepwork`:
|
|
|
133
133
|
- **Execution Flow:** Mermaid diagram (3-4 nodes).
|
|
134
134
|
- **Slice Table:** Compact table (`Phase | Specialist | Key Deliverable | Verification Slice`).
|
|
135
135
|
- **Disk Blueprint:** Clear reference that the exhaustive technical spec is preserved in `.pantheon/deepwork/<slug>/PLAN.md`.
|
|
136
|
-
5. **Interactive Question Gate
|
|
136
|
+
5. **Interactive Question Gate (MANDATORY `question` tool):**
|
|
137
|
+
- **Open Questions & Decisions Needed:** If the plan, task brief, or refinement contains any Open Questions or `decisions_needed`, **NEVER leave them as inert text bullets** in chat or assume "proceed if no objections". You MUST decompose each open question into an interactive item in the native `question` tool with concrete options, marking the recommended choice with `(Recomendado)`.
|
|
138
|
+
- **Plan Approval Gate:** Alongside any open questions, include the approval question with explicit options (e.g., `["(Recomendado) Aprovar e iniciar execução", "Ajustar plano"]` / `["(Recommended) Approve and begin execution", "Adjust plan"]`), falling back to plain text only if the tool is not exposed.
|
|
137
139
|
6. If the user requests changes, send feedback to Athena, update `.pantheon/deepwork/<slug>/PLAN.md`, and repeat the interactive gate.
|
|
@@ -22,7 +22,7 @@ For investigation agents (Apollo, Athena), return structured findings: key findi
|
|
|
22
22
|
|
|
23
23
|
### Athena Plan Handoff
|
|
24
24
|
|
|
25
|
-
When Zeus delegates Athena through the root `subagent` dispatch, its final task result MUST contain the complete Markdown plan (not a summary, artifact path, or approval question): goal, scope and assumptions, phases, affected files or areas, dependencies, verification, risks, open questions, acceptance criteria — unknowns marked, never guessed. This supersedes the structured-findings format. Athena must not call `agent/askQuestions`, pause for user input, or ask for approval in the child session. Zeus publishes the full plan
|
|
25
|
+
When Zeus delegates Athena through the root `subagent` dispatch, its final task result MUST contain the complete Markdown plan (not a summary, artifact path, or approval question): goal, scope and assumptions, phases, affected files or areas, dependencies, verification, risks, open questions, acceptance criteria — unknowns marked, never guessed. This supersedes the structured-findings format. Athena must not call `agent/askQuestions`, pause for user input, or ask for approval in the child session. Zeus publishes the full plan before approval UI, never bundling them; if order is not guaranteed, Zeus asks in plain text afterwards. For Open Questions or decisions needed, Zeus prompts via the native `question` tool with structured options and marked recommended defaults, never leaving inert text bullets. Direct Athena invocations show the complete plan before requesting approval.
|
|
26
26
|
|
|
27
27
|
## Council Specialist Response Format
|
|
28
28
|
|
|
@@ -153,7 +153,7 @@ export function buildV2VisionPromptContext(
|
|
|
153
153
|
payload.push(
|
|
154
154
|
'call vision_describe with the image path; handoff @aphrodite; an image never authorizes a revert',
|
|
155
155
|
)
|
|
156
|
-
return ['<!-- pantheon-
|
|
156
|
+
return ['<!-- pantheon-visual-context -->', payload.join(' · ')].join('\n')
|
|
157
157
|
}
|
|
158
158
|
|
|
159
159
|
export interface V2VisionModel {
|
package/src/plugin/index.js
CHANGED
|
@@ -8780,7 +8780,7 @@ function buildV2VisionPromptContext(files, userText, sessionID) {
|
|
|
8780
8780
|
payload.push(
|
|
8781
8781
|
"call vision_describe with the image path; handoff @aphrodite; an image never authorizes a revert"
|
|
8782
8782
|
);
|
|
8783
|
-
return ["<!-- pantheon-
|
|
8783
|
+
return ["<!-- pantheon-visual-context -->", payload.join(" \xB7 ")].join("\n");
|
|
8784
8784
|
}
|
|
8785
8785
|
function asRecord2(value) {
|
|
8786
8786
|
return typeof value === "object" && value !== null && !Array.isArray(value) ? value : void 0;
|
|
@@ -14244,7 +14244,7 @@ var UNSUPPORTED_FEATURE_SEED = Object.freeze([
|
|
|
14244
14244
|
|
|
14245
14245
|
// src/plugin.ts
|
|
14246
14246
|
var V2_UNSUPPORTED_FEATURES = [...UNSUPPORTED_FEATURE_SEED];
|
|
14247
|
-
var POLICY_MARKER = "<!-- pantheon-
|
|
14247
|
+
var POLICY_MARKER = "<!-- pantheon-policy -->";
|
|
14248
14248
|
var POLICY = `${POLICY_MARKER}
|
|
14249
14249
|
Follow Pantheon routing policy: delegate implementation work to the named specialist and do not claim work was performed without verification.`;
|
|
14250
14250
|
function isTextPart(entry) {
|
|
@@ -14792,7 +14792,7 @@ async function registerV2SessionHooks(context) {
|
|
|
14792
14792
|
const prompt = admission?.prompt;
|
|
14793
14793
|
if (!prompt || typeof prompt.text !== "string") return;
|
|
14794
14794
|
const userRequest = prompt.text;
|
|
14795
|
-
if (!prompt.text.includes("<!-- pantheon-
|
|
14795
|
+
if (!prompt.text.includes("<!-- pantheon-visual-context -->")) {
|
|
14796
14796
|
const sessionID = typeof admission?.sessionID === "string" && admission.sessionID.trim() !== "" ? admission.sessionID : void 0;
|
|
14797
14797
|
const nativeByLookup = await (async () => {
|
|
14798
14798
|
if (sessionID === void 0) return false;
|
package/src/plugin.ts
CHANGED
|
@@ -211,7 +211,7 @@ export const V2_UNSUPPORTED_FEATURES: string[] = [...UNSUPPORTED_FEATURE_SEED]
|
|
|
211
211
|
|
|
212
212
|
// ─── Constants ───────────────────────────────────────────────────────────
|
|
213
213
|
|
|
214
|
-
const POLICY_MARKER = '<!-- pantheon-
|
|
214
|
+
const POLICY_MARKER = '<!-- pantheon-policy -->'
|
|
215
215
|
const POLICY = `${POLICY_MARKER}\nFollow Pantheon routing policy: delegate implementation work to the named specialist and do not claim work was performed without verification.`
|
|
216
216
|
|
|
217
217
|
/** 2.0.16 `system` on a session hook payload is `Array<SystemPart>`. */
|
|
@@ -1460,7 +1460,7 @@ async function registerV2SessionHooks(context: PluginContext): Promise<boolean>
|
|
|
1460
1460
|
const prompt = admission?.prompt
|
|
1461
1461
|
if (!prompt || typeof prompt.text !== 'string') return
|
|
1462
1462
|
const userRequest = prompt.text
|
|
1463
|
-
if (!prompt.text.includes('<!-- pantheon-
|
|
1463
|
+
if (!prompt.text.includes('<!-- pantheon-visual-context -->')) {
|
|
1464
1464
|
const sessionID =
|
|
1465
1465
|
typeof admission?.sessionID === 'string' && admission.sessionID.trim() !== ''
|
|
1466
1466
|
? admission.sessionID
|
|
@@ -2062,7 +2062,7 @@ async function readChangelogForBundle(bundleUrl) {
|
|
|
2062
2062
|
}
|
|
2063
2063
|
/** The ambient build-time version, or null outside a tsdown build. */
|
|
2064
2064
|
function embeddedVersion() {
|
|
2065
|
-
return "2.1.
|
|
2065
|
+
return "2.1.2";
|
|
2066
2066
|
}
|
|
2067
2067
|
/**
|
|
2068
2068
|
* Resolve the running Pantheon version for the what's-new window: the plugin
|
|
@@ -2287,7 +2287,7 @@ function persistUpdateCheckState(previous, latest) {
|
|
|
2287
2287
|
} catch {}
|
|
2288
2288
|
}
|
|
2289
2289
|
function resolveRunningVersion() {
|
|
2290
|
-
return "2.1.
|
|
2290
|
+
return "2.1.2";
|
|
2291
2291
|
}
|
|
2292
2292
|
const VALUE_WIDTH = 24;
|
|
2293
2293
|
function SettingsWindow(props) {
|
|
@@ -6513,7 +6513,7 @@ async function checkForPantheonUpdate(context) {
|
|
|
6513
6513
|
const latest = String((await response.json())?.version ?? "");
|
|
6514
6514
|
if (!latest) return;
|
|
6515
6515
|
const notice = pickUpdateNotice({
|
|
6516
|
-
currentVersion: "2.1.
|
|
6516
|
+
currentVersion: "2.1.2",
|
|
6517
6517
|
latestVersion: latest,
|
|
6518
6518
|
state: readState(),
|
|
6519
6519
|
nowMs: Date.now()
|
|
@@ -7325,7 +7325,7 @@ function Sidebar(props) {
|
|
|
7325
7325
|
get children() {
|
|
7326
7326
|
var _el$13 = createElement("text");
|
|
7327
7327
|
setProp(_el$13, "attributes", 1);
|
|
7328
|
-
insert(_el$13, `Pantheon v2.1.
|
|
7328
|
+
insert(_el$13, `Pantheon v2.1.2`);
|
|
7329
7329
|
effect((_$p) => setProp(_el$13, "fg", text(), _$p));
|
|
7330
7330
|
return _el$13;
|
|
7331
7331
|
}
|
|
@@ -8097,7 +8097,7 @@ function HomeStatus(props) {
|
|
|
8097
8097
|
},
|
|
8098
8098
|
get children() {
|
|
8099
8099
|
var _el$73 = createElement("text");
|
|
8100
|
-
insert(_el$73, () => `Pantheon v2.1.
|
|
8100
|
+
insert(_el$73, () => `Pantheon v2.1.2${preset().name ? ` ▸ ${preset().name}` : ""}`);
|
|
8101
8101
|
effect((_$p) => setProp(_el$73, "fg", muted(), _$p));
|
|
8102
8102
|
return _el$73;
|
|
8103
8103
|
}
|
|
@@ -1,12 +1,12 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "pantheon-tui",
|
|
3
|
-
"version": "2.1.
|
|
3
|
+
"version": "2.1.2",
|
|
4
4
|
"lockfileVersion": 3,
|
|
5
5
|
"requires": true,
|
|
6
6
|
"packages": {
|
|
7
7
|
"": {
|
|
8
8
|
"name": "pantheon-tui",
|
|
9
|
-
"version": "2.1.
|
|
9
|
+
"version": "2.1.2",
|
|
10
10
|
"license": "MIT",
|
|
11
11
|
"dependencies": {
|
|
12
12
|
"@opencode/plugin": "2.0.26",
|
|
@@ -59,7 +59,7 @@ Presented to user for confirmation before starting:
|
|
|
59
59
|
- [ ] Observable completion condition
|
|
60
60
|
- [ ] Test verification passes cleanly
|
|
61
61
|
- **Open Questions (with Recommended Defaults):**
|
|
62
|
-
- [Question]? *(Default: [Recommended choice] —
|
|
62
|
+
- [Question]? *(Default: [Recommended choice] — interactively confirmed via question tool)*
|
|
63
63
|
```
|
|
64
64
|
|
|
65
65
|
---
|
|
@@ -70,7 +70,9 @@ Presented to user for confirmation before starting:
|
|
|
70
70
|
2. **Tier 2 Requires `>>`:** Only explicit `>>` prefix requests trigger user-visible contract confirmation.
|
|
71
71
|
3. **Language Consistency:** Always preserve the user's language (PT-BR requests produce PT-BR briefs).
|
|
72
72
|
4. **Fail-Open:** Any error during refinement falls back to the original prompt without blocking.
|
|
73
|
-
5. **Interactive Questioning & Ambiguity Resolution:**
|
|
74
|
-
- Whenever prompt refinement
|
|
73
|
+
5. **Interactive Questioning & Ambiguity Resolution (MANDATORY `question` tool):**
|
|
74
|
+
- Whenever prompt refinement or planning surfaces Open Questions (`decisions_needed`), trade-offs, or scope choices, **NEVER leave them as inert text bullets in chat** or assume "proceed if no objections".
|
|
75
|
+
- You **MUST ALWAYS invoke the native `question` tool** with structured choices, placing the recommended choice first and marking it with `(Recomendado)`.
|
|
76
|
+
- The user selects choices with 1 click/Enter, and the selections are recorded as audited decisions.
|
|
75
77
|
- With decision router active, use `pantheon_jev_decide` (mode: `choice`) for internal candidate scoring.
|
|
76
|
-
- Non-blocking minor gaps use sensible declared defaults without stalling the flow.
|
|
78
|
+
- Non-blocking minor gaps without meaningful user trade-offs use sensible declared defaults without stalling the flow.
|