teamcast 1.0.0 → 1.0.1
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/README.md +37 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -432,6 +432,9 @@ It targets:
|
|
|
432
432
|
- `.claude/settings.json`
|
|
433
433
|
- `.claude/settings.local.json`
|
|
434
434
|
- `CLAUDE.md`
|
|
435
|
+
- `.codex/agents`
|
|
436
|
+
- `.codex/config.toml`
|
|
437
|
+
- `AGENTS.md`
|
|
435
438
|
|
|
436
439
|
Flow:
|
|
437
440
|
|
|
@@ -642,6 +645,7 @@ claude:
|
|
|
642
645
|
| `claude.max_turns` | number | Maximum agentic turns |
|
|
643
646
|
| `claude.mcp_servers` | object[] | MCP server definitions |
|
|
644
647
|
| `claude.permission_mode` | `default \| acceptEdits \| bypassPermissions \| plan \| dontAsk` | Claude Code permission mode |
|
|
648
|
+
| `capability_traits` | string[] | High-level capability profiles that expand to `tools` + `disallowed_tools` |
|
|
645
649
|
| `instruction_blocks` | object[] | Structured instruction blocks rendered into Claude markdown and Codex `developer_instructions` |
|
|
646
650
|
| `instruction_fragments` | string[] | Named instruction fragments composed with `instruction_blocks` |
|
|
647
651
|
| `forge.handoffs` | string[] | Other agents this agent can delegate to. Requires `Agent` in `claude.tools` |
|
|
@@ -675,6 +679,39 @@ developer:
|
|
|
675
679
|
|
|
676
680
|
Raw tool names and skills can be mixed in the same `tools` array.
|
|
677
681
|
|
|
682
|
+
### Capability Traits
|
|
683
|
+
|
|
684
|
+
Capability traits are high-level profiles that compose `tools` and `disallowed_tools` from named capability sets. They are the recommended way to define agent permissions in presets and custom manifests — more readable than raw tool lists and safer than manual assembly.
|
|
685
|
+
|
|
686
|
+
| Trait | Grants | Denies |
|
|
687
|
+
| -------------------- | ------------------------------- | --------------- |
|
|
688
|
+
| `base-read` | `read_files`, `search` | — |
|
|
689
|
+
| `file-authoring` | `write_files` | — |
|
|
690
|
+
| `command-execution` | `execute` | — |
|
|
691
|
+
| `web-research` | `web` | — |
|
|
692
|
+
| `delegation` | `delegate` | — |
|
|
693
|
+
| `interaction` | `interact` | — |
|
|
694
|
+
| `notebook-editing` | `notebook` | — |
|
|
695
|
+
| `no-file-edits` | — | `write_files` |
|
|
696
|
+
| `no-commands` | — | `execute` |
|
|
697
|
+
| `no-web` | — | `web` |
|
|
698
|
+
| `full-access` | all capabilities | — |
|
|
699
|
+
|
|
700
|
+
Grant traits expand to the corresponding abstract skills (see table above). Deny traits add the expanded tools to `disallowed_tools`.
|
|
701
|
+
|
|
702
|
+
Traits are combined in order. If you need a read-only planner that can browse the web but cannot edit files or run commands:
|
|
703
|
+
|
|
704
|
+
```yaml
|
|
705
|
+
planner:
|
|
706
|
+
capability_traits:
|
|
707
|
+
- base-read
|
|
708
|
+
- web-research
|
|
709
|
+
- no-file-edits
|
|
710
|
+
- no-commands
|
|
711
|
+
```
|
|
712
|
+
|
|
713
|
+
Traits merge with explicit `tools` and `disallowed_tools` — you can use both in the same agent definition. Explicit tool lists take precedence for additions; deny traits always add to the disallow list.
|
|
714
|
+
|
|
678
715
|
### Policy Fields
|
|
679
716
|
|
|
680
717
|
```yaml
|