@cubis/foundry 0.3.72 → 0.3.73
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/CHANGELOG.md +8 -2
- package/dist/cli/core.js +5 -4
- package/dist/cli/core.js.map +1 -1
- package/package.json +1 -1
- package/src/cli/core.ts +9 -4
- package/workflows/workflows/agent-environment-setup/platforms/antigravity/rules/GEMINI.md +65 -42
- package/workflows/workflows/agent-environment-setup/platforms/claude/rules/CLAUDE.md +8 -6
- package/workflows/workflows/agent-environment-setup/platforms/codex/rules/AGENTS.md +65 -41
- package/workflows/workflows/agent-environment-setup/platforms/copilot/rules/copilot-instructions.md +8 -6
- package/workflows/workflows/agent-environment-setup/shared/rules/STEERING.md +9 -8
- package/workflows/workflows/agent-environment-setup/shared/rules/overrides/codex.md +1 -1
|
@@ -47,16 +47,17 @@ Follow this decision tree for EVERY user request:
|
|
|
47
47
|
|
|
48
48
|
## 3) Skill Loading Protocol
|
|
49
49
|
|
|
50
|
-
Skills are **supporting context
|
|
50
|
+
Skills are **supporting context** — always route first, then load what the route recommends.
|
|
51
51
|
|
|
52
52
|
1. **Never begin with `skill_search`.** Inspect the repo/task locally first.
|
|
53
|
-
2. Resolve the route (workflow, agent, or direct execution) before
|
|
54
|
-
3.
|
|
55
|
-
4. If
|
|
56
|
-
5.
|
|
57
|
-
6.
|
|
58
|
-
7.
|
|
59
|
-
8.
|
|
53
|
+
2. Resolve the route (workflow, agent, or direct execution) before loading any skills.
|
|
54
|
+
3. **After routing: if `route_resolve` returned `primarySkillHint` or `primarySkills`, load the first via `skill_validate` → `skill_get` before executing. Not optional for non-trivial tasks.**
|
|
55
|
+
4. If `detectedLanguageSkill` is returned and matches the project, load it too (if not already loaded this session).
|
|
56
|
+
5. Domain still unclear after routing? → ONE narrow `skill_search`. Not two.
|
|
57
|
+
6. Call `skill_get` with `includeReferences: false` by default.
|
|
58
|
+
7. Load reference files one at a time with `skill_get_reference` — only when a specific reference is needed.
|
|
59
|
+
8. Do not auto-prime every specialist. Only load what `primarySkills` recommends or the task clearly needs.
|
|
60
|
+
9. Never pass workflow IDs or agent IDs to skill tools.
|
|
60
61
|
|
|
61
62
|
---
|
|
62
63
|
|
|
@@ -19,7 +19,7 @@
|
|
|
19
19
|
- All specialist references are **postures within the current session** — Codex does not spawn isolated subagents.
|
|
20
20
|
- `@specialist` means: adopt that specialist's domain, reasoning style, and scope constraints internally.
|
|
21
21
|
- Prefer local file inspection over external fetches. Default to repo-grounded reasoning.
|
|
22
|
-
- MCP tools
|
|
22
|
+
- Foundry MCP tools (`skill_get`, `skill_search`, `skill_validate`, `route_resolve`) are available when the MCP server is connected. After `route_resolve`, load the returned `primarySkillHint` or `primarySkills[0]` via `skill_validate` → `skill_get` before executing non-trivial tasks.
|
|
23
23
|
|
|
24
24
|
## Platform Notes
|
|
25
25
|
|