@team-agent/installer 0.5.125 → 0.5.126
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/Cargo.lock +1 -1
- package/Cargo.toml +1 -1
- package/crates/team-agent/src/cli/diagnose.rs +0 -68
- package/crates/team-agent/src/cli/emit.rs +6 -2
- package/crates/team-agent/src/cli/intent/runtime_tests.rs +10 -10
- package/crates/team-agent/src/cli/mod.rs +22 -28
- package/crates/team-agent/src/cli/models.rs +7 -764
- package/crates/team-agent/src/cli/tests/mod.rs +1 -1
- package/crates/team-agent/src/cli/tests/run_delegation.rs +5 -4
- package/crates/team-agent/src/compiler/tests.rs +18 -3
- package/crates/team-agent/src/compiler.rs +4 -203
- package/crates/team-agent/src/contract_runtime/backend.rs +61 -38
- package/crates/team-agent/src/contract_runtime/config.rs +7 -4
- package/crates/team-agent/src/contract_runtime/fork.rs +368 -0
- package/crates/team-agent/src/contract_runtime/framework.rs +8 -10
- package/crates/team-agent/src/contract_runtime/mod.rs +1 -0
- package/crates/team-agent/src/contract_runtime/registry.rs +60 -19
- package/crates/team-agent/src/coordinator/tests/api_error_recovery.rs +29 -8
- package/crates/team-agent/src/coordinator/tests/health_sync.rs +0 -42
- package/crates/team-agent/src/leader/start.rs +1 -1
- package/crates/team-agent/src/lifecycle/launch/add_agent_state.rs +2 -195
- package/crates/team-agent/src/lifecycle/launch/clone_agent.rs +11 -27
- package/crates/team-agent/src/lifecycle/launch/fork_agent.rs +18 -28
- package/crates/team-agent/src/lifecycle/launch/fork_entry.rs +6 -87
- package/crates/team-agent/src/lifecycle/launch/mcp_config.rs +1 -9
- package/crates/team-agent/src/lifecycle/launch/pi_mcp.rs +74 -244
- package/crates/team-agent/src/lifecycle/launch/quick_start.rs +0 -24
- package/crates/team-agent/src/lifecycle/launch/role_source.rs +5 -0
- package/crates/team-agent/src/lifecycle/launch/spawn.rs +21 -33
- package/crates/team-agent/src/lifecycle/launch.rs +0 -23
- package/crates/team-agent/src/lifecycle/profile_launch.rs +18 -317
- package/crates/team-agent/src/lifecycle/restart/common.rs +42 -315
- package/crates/team-agent/src/lifecycle/restart/selection.rs +5 -5
- package/crates/team-agent/src/lifecycle/restart.rs +0 -2
- package/crates/team-agent/src/lifecycle/tests/behavioral_diff_264_red.rs +11 -6
- package/crates/team-agent/src/lifecycle/tests/claude_profile_launch_red.rs +8 -320
- package/crates/team-agent/src/lifecycle/tests/cli_argv_routing_red.rs +22 -138
- package/crates/team-agent/src/lifecycle/tests/clone_fork_copilot_perms_red.rs +4 -334
- package/crates/team-agent/src/lifecycle/tests/communication_mode_runtime_contract_red.rs +30 -11
- package/crates/team-agent/src/lifecycle/tests/copilot_provider_red.rs +2 -2
- package/crates/team-agent/src/lifecycle/tests/core.rs +17 -16
- package/crates/team-agent/src/lifecycle/tests/grok_require_explicit_model_red.rs +12 -12
- package/crates/team-agent/src/lifecycle/tests/harvest2_a_batch_red.rs +3 -51
- package/crates/team-agent/src/lifecycle/tests/launch_spawn.rs +26 -28
- package/crates/team-agent/src/lifecycle/tests/pi_catalog_red.rs +6 -18
- package/crates/team-agent/src/lifecycle/tests/pi_command_red.rs +52 -70
- package/crates/team-agent/src/lifecycle/tests/pi_dual_entry_red.rs +4 -157
- package/crates/team-agent/src/lifecycle/tests/pi_executable_mcp_red.rs +10 -302
- package/crates/team-agent/src/lifecycle/tests/pi_lifecycle_doctor_red.rs +1 -68
- package/crates/team-agent/src/lifecycle/tests/pi_native_mcp_capability_red.rs +1 -19
- package/crates/team-agent/src/lifecycle/tests/pi_session_delivery_red.rs +2 -100
- package/crates/team-agent/src/lifecycle/tests/quick_start_worker_readiness_red.rs +11 -10
- package/crates/team-agent/src/lifecycle/tests/restart_session_capture_red.rs +2 -106
- package/crates/team-agent/src/lifecycle/tests/status_credential_redaction_contract.rs +2 -2
- package/crates/team-agent/src/lifecycle/tests/swallow_batch4_semantics_red.rs +2 -254
- package/crates/team-agent/src/lifecycle/tests/upgrade_compat_0211_red.rs +4 -4
- package/crates/team-agent/src/lifecycle/tests/verify_rs031_window_consistency_red.rs +6 -5
- package/crates/team-agent/src/lifecycle/tests/worker_spawn_env_red.rs +7 -16
- package/crates/team-agent/src/lifecycle/tests.rs +0 -3
- package/crates/team-agent/src/lifecycle/types.rs +2 -18
- package/crates/team-agent/src/mcp_server/readiness_tests.rs +1 -1
- package/crates/team-agent/src/messaging/delivery.rs +12 -141
- package/crates/team-agent/src/messaging/delivery_consumption_tests.rs +5 -5
- package/crates/team-agent/src/messaging/tests/dup_inject.rs +3 -3
- package/crates/team-agent/src/messaging/tests/spine.rs +26 -32
- package/crates/team-agent/src/provider/adapter.rs +35 -238
- package/crates/team-agent/src/provider/adapters/mod.rs +4 -4
- package/crates/team-agent/src/provider/model_catalog.rs +13 -201
- package/crates/team-agent/src/provider/session/capture.rs +6 -106
- package/crates/team-agent/src/provider/session/context_fork/outcome.rs +0 -4
- package/crates/team-agent/src/provider/session/context_fork.rs +15 -107
- package/crates/team-agent/src/provider/session_scan/common.rs +10 -34
- package/crates/team-agent/src/provider/session_scan/pi.rs +1 -34
- package/crates/team-agent/src/provider/session_scan.rs +4 -10
- package/crates/team-agent/src/provider/tests/adapter.rs +34 -91
- package/crates/team-agent/src/provider/tests/command_helpers.rs +3 -12
- package/crates/team-agent/src/provider/tests/idle.rs +0 -25
- package/crates/team-agent/src/provider/wire.rs +51 -23
- package/crates/team-agent/src/state/abnormal_watch/tests/mod.rs +1 -1
- package/crates/team-agent/src/state/persist.rs +6 -6
- package/crates/team-agent/src/state/repository/tests.rs +2 -2
- package/package.json +4 -4
- package/skills/team-agent/SKILL.md +6 -4
- package/skills/team-agent/command-coverage.json +15 -1
- package/skills/team-agent/docs/reference/team-agent-operator.md +5 -3
- package/skills/team-agent/references/team-agent-operator.md +5 -3
- package/crates/team-agent/src/lifecycle/restart/avatar.rs +0 -1805
- package/crates/team-agent/src/lifecycle/tests/claude_compatible_config_red.rs +0 -325
- package/crates/team-agent/src/lifecycle/tests/pi_compiler_red.rs +0 -210
- package/crates/team-agent/src/lifecycle/tests/pi_fork_session_inheritance_red.rs +0 -1409
- package/crates/team-agent/src/provider/adapters/claude.rs +0 -78
- package/crates/team-agent/src/provider/adapters/claude_fork.rs +0 -131
- package/crates/team-agent/src/provider/adapters/pi.rs +0 -100
- package/crates/team-agent/src/provider/session/context_fork/claude.rs +0 -123
- package/crates/team-agent/src/provider/session_scan/claude.rs +0 -523
- package/crates/team-agent/src/test_support/models_timeout_fixture.rs +0 -144
|
@@ -2730,7 +2730,7 @@ mod tests {
|
|
|
2730
2730
|
&json!({
|
|
2731
2731
|
"session_name": "team-a",
|
|
2732
2732
|
"active_team_key": "team-a",
|
|
2733
|
-
"agents": {"f1": {"provider": "
|
|
2733
|
+
"agents": {"f1": {"provider": "copilot", "status": "starting"}}
|
|
2734
2734
|
}),
|
|
2735
2735
|
);
|
|
2736
2736
|
// f2 loaded before f1's reservation. Persist carries the still-unspawned
|
|
@@ -2738,7 +2738,7 @@ mod tests {
|
|
|
2738
2738
|
let incoming = json!({
|
|
2739
2739
|
"session_name": "team-a",
|
|
2740
2740
|
"active_team_key": "team-a",
|
|
2741
|
-
"agents": {"f2": {"provider": "
|
|
2741
|
+
"agents": {"f2": {"provider": "copilot", "status": "starting"}}
|
|
2742
2742
|
});
|
|
2743
2743
|
save_runtime_state_with_lifecycle_topology_authority(&ws, &incoming, "team-a", &["f2"])
|
|
2744
2744
|
.unwrap();
|
|
@@ -2749,7 +2749,7 @@ mod tests {
|
|
|
2749
2749
|
// f1 starts after f2 takes its snapshot; capture is not complete yet.
|
|
2750
2750
|
let mut started = stale.clone();
|
|
2751
2751
|
started["agents"]["f1"] = json!({
|
|
2752
|
-
"provider": "
|
|
2752
|
+
"provider": "copilot",
|
|
2753
2753
|
"status": "running",
|
|
2754
2754
|
"pane_id": "%1",
|
|
2755
2755
|
"pane_pid": 101,
|
|
@@ -2763,7 +2763,7 @@ mod tests {
|
|
|
2763
2763
|
.unwrap();
|
|
2764
2764
|
let latest = read_state(&ws);
|
|
2765
2765
|
stale["agents"]["f2"] = json!({
|
|
2766
|
-
"provider": "
|
|
2766
|
+
"provider": "copilot",
|
|
2767
2767
|
"status": "running",
|
|
2768
2768
|
"pane_id": "%2",
|
|
2769
2769
|
"pane_pid": 102,
|
|
@@ -3410,7 +3410,7 @@ mod tests {
|
|
|
3410
3410
|
"agents": {
|
|
3411
3411
|
"alpha": {
|
|
3412
3412
|
"status": "stopped",
|
|
3413
|
-
"provider": "
|
|
3413
|
+
"provider": "copilot",
|
|
3414
3414
|
"session_id": "stale-poisoned-uuid",
|
|
3415
3415
|
"rollout_path": Value::Null,
|
|
3416
3416
|
"captured_at": Value::Null,
|
|
@@ -3427,7 +3427,7 @@ mod tests {
|
|
|
3427
3427
|
"agents": {
|
|
3428
3428
|
"alpha": {
|
|
3429
3429
|
"status": "running",
|
|
3430
|
-
"provider": "
|
|
3430
|
+
"provider": "copilot",
|
|
3431
3431
|
"session_id": Value::Null,
|
|
3432
3432
|
"rollout_path": Value::Null,
|
|
3433
3433
|
"captured_at": Value::Null,
|
|
@@ -123,9 +123,9 @@ fn stopped_start_commit_updates_exact_target_and_both_mirrors()
|
|
|
123
123
|
let held = json!({"agent_id":"worker","provider":"grok_build","runtime_family":"contract",
|
|
124
124
|
"status":"stopped","explicit_stop":true,"generation":1,"contract_instance":"instance-one",
|
|
125
125
|
"session_id":"same-session","native_endpoint":"old-endpoint"});
|
|
126
|
-
let peer = json!({"agent_id":"peer","provider":"
|
|
126
|
+
let peer = json!({"agent_id":"peer","provider":"copilot","status":"stopped","explicit_stop":true,
|
|
127
127
|
"session_id":"peer-session","generation":1});
|
|
128
|
-
let control = json!({"agent_id":"control","provider":"
|
|
128
|
+
let control = json!({"agent_id":"control","provider":"copilot","status":"running","session_id":"control-session"});
|
|
129
129
|
let agents = json!({"worker":held,"peer":peer,"control":control});
|
|
130
130
|
let initial = json!({"session_name":"team-a","active_team_key":"team-a","agents":agents,
|
|
131
131
|
"teams":{"team-a":{"session_name":"team-a","status":"alive","agents":agents},
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@team-agent/installer",
|
|
3
|
-
"version": "0.5.
|
|
3
|
+
"version": "0.5.126",
|
|
4
4
|
"description": "npx installer for Team Agent",
|
|
5
5
|
"keywords": [
|
|
6
6
|
"codex",
|
|
@@ -20,9 +20,9 @@
|
|
|
20
20
|
"team-agent-installer": "npm/install.mjs"
|
|
21
21
|
},
|
|
22
22
|
"optionalDependencies": {
|
|
23
|
-
"@team-agent/cli-darwin-arm64": "0.5.
|
|
24
|
-
"@team-agent/cli-darwin-x64": "0.5.
|
|
25
|
-
"@team-agent/cli-linux-x64": "0.5.
|
|
23
|
+
"@team-agent/cli-darwin-arm64": "0.5.126",
|
|
24
|
+
"@team-agent/cli-darwin-x64": "0.5.126",
|
|
25
|
+
"@team-agent/cli-linux-x64": "0.5.126"
|
|
26
26
|
},
|
|
27
27
|
"scripts": {
|
|
28
28
|
"postinstall": "node npm/bincheck.mjs",
|
|
@@ -19,8 +19,9 @@ Use existing resources rather than establishing accounts. `team-agent doctor --w
|
|
|
19
19
|
|
|
20
20
|
| Choice | Explain |
|
|
21
21
|
|---|---|
|
|
22
|
-
| Pi | Several model catalogs;
|
|
23
|
-
|
|
|
22
|
+
| Pi | Several model catalogs; every Pi worker needs an exact qualified model ID such as `openai-codex/gpt-6-luna` from `team-agent models --provider pi`; Pi has no permission prompt to bypass, so keep bypass `false` |
|
|
23
|
+
| Claude | Every Claude worker needs an exact model ID from `team-agent models --provider claude`; existing subscription login only |
|
|
24
|
+
| Codex / other native CLIs | The user's existing workflow, native permissions, model catalog and supported session capabilities |
|
|
24
25
|
| Bypass `true` | Fewer interruptions, broader automatic authority |
|
|
25
26
|
| Bypass `false` | Native permissions remain; human confirmation may interrupt unattended work |
|
|
26
27
|
|
|
@@ -36,11 +37,12 @@ Keep the goal in `TEAM.md` and each worker's instructions in `agents/*.md`. The
|
|
|
36
37
|
agents/reviewer.md
|
|
37
38
|
```
|
|
38
39
|
|
|
39
|
-
Record provider and boolean bypass in every new role. This example assumes the user explicitly chose Pi and bypass false; it is not a default:
|
|
40
|
+
Record provider, model and boolean bypass in every new role. This example assumes the user explicitly chose Pi, a model copied from `team-agent models --provider pi`, and bypass false; it is not a default:
|
|
40
41
|
|
|
41
42
|
```markdown
|
|
42
43
|
---
|
|
43
44
|
provider: pi
|
|
45
|
+
model: openai-codex/gpt-6-luna
|
|
44
46
|
dangerously_skip_permissions: false
|
|
45
47
|
---
|
|
46
48
|
Review the change and report concrete regression risks.
|
|
@@ -88,7 +90,7 @@ With the user's provider/bypass choices, `start` unifies adding an absent seat a
|
|
|
88
90
|
|
|
89
91
|
```sh
|
|
90
92
|
team-agent start auditor --role-file /absolute/path/to/auditor.md
|
|
91
|
-
team-agent start analyst --provider pi --bypass false --prompt "Analyze the proposed change"
|
|
93
|
+
team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt "Analyze the proposed change"
|
|
92
94
|
team-agent stop reviewer
|
|
93
95
|
team-agent start reviewer --effort high --bypass false
|
|
94
96
|
```
|
|
@@ -490,6 +490,20 @@
|
|
|
490
490
|
"owner": "runtime-owner:public-cli",
|
|
491
491
|
"plan": "Bind this exact documented argv to an isolated HermeticTestEnv result and an assertion/negative-twin pair before promotion. For lifecycle use an authorized owned team and verify scope, identity, outcome and cleanup; help or readonly plans alone do not establish lifecycle coverage."
|
|
492
492
|
},
|
|
493
|
+
{
|
|
494
|
+
"command": "team-agent models --provider claude",
|
|
495
|
+
"bucket": "declared_gap",
|
|
496
|
+
"covered": false,
|
|
497
|
+
"owner": "runtime-owner:public-cli",
|
|
498
|
+
"plan": "Bind this exact documented argv to an isolated HermeticTestEnv result and an assertion/negative-twin pair before promotion. For lifecycle use an authorized owned team and verify scope, identity, outcome and cleanup; help or readonly plans alone do not establish lifecycle coverage."
|
|
499
|
+
},
|
|
500
|
+
{
|
|
501
|
+
"command": "team-agent models --provider pi",
|
|
502
|
+
"bucket": "declared_gap",
|
|
503
|
+
"covered": false,
|
|
504
|
+
"owner": "runtime-owner:public-cli",
|
|
505
|
+
"plan": "Bind this exact documented argv to an isolated HermeticTestEnv result and an assertion/negative-twin pair before promotion. For lifecycle use an authorized owned team and verify scope, identity, outcome and cleanup; help or readonly plans alone do not establish lifecycle coverage."
|
|
506
|
+
},
|
|
493
507
|
{
|
|
494
508
|
"command": "team-agent profile init my-api --auth-mode compatible_api --workspace .",
|
|
495
509
|
"bucket": "declared_gap",
|
|
@@ -596,7 +610,7 @@
|
|
|
596
610
|
"plan": "Bind this exact documented argv to an isolated HermeticTestEnv result and an assertion/negative-twin pair before promotion. For lifecycle use an authorized owned team and verify scope, identity, outcome and cleanup; help or readonly plans alone do not establish lifecycle coverage."
|
|
597
611
|
},
|
|
598
612
|
{
|
|
599
|
-
"command": "team-agent start analyst --provider pi --bypass false --prompt \"Analyze the proposed change\"",
|
|
613
|
+
"command": "team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt \"Analyze the proposed change\"",
|
|
600
614
|
"bucket": "declared_gap",
|
|
601
615
|
"covered": false,
|
|
602
616
|
"owner": "runtime-owner:public-cli",
|
|
@@ -18,6 +18,8 @@ team-agent agent restart reviewer --session preserve --workspace . --team projec
|
|
|
18
18
|
team-agent doctor --workspace . --json
|
|
19
19
|
team-agent inbox leader --workspace . --team project-team -n 3
|
|
20
20
|
team-agent models --provider NAME
|
|
21
|
+
team-agent models --provider claude
|
|
22
|
+
team-agent models --provider pi
|
|
21
23
|
team-agent profile init my-api --auth-mode compatible_api --workspace .
|
|
22
24
|
team-agent restart reviewer --workspace . --team project-team
|
|
23
25
|
team-agent route add pi -- --verbose
|
|
@@ -33,7 +35,7 @@ team-agent start --from ./my-team --workspace . --team project-team
|
|
|
33
35
|
team-agent start --from .team/current --workspace . --team project-team
|
|
34
36
|
team-agent start --help
|
|
35
37
|
team-agent start --workspace . --team project-team
|
|
36
|
-
team-agent start analyst --provider pi --bypass false --prompt "Analyze the proposed change"
|
|
38
|
+
team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt "Analyze the proposed change"
|
|
37
39
|
team-agent start auditor --role-file /absolute/path/to/auditor.md
|
|
38
40
|
team-agent start reviewer --effort high --bypass false
|
|
39
41
|
team-agent start reviewer --effort high --bypass false --workspace . --team project-team
|
|
@@ -140,7 +142,7 @@ Creation needs complete configuration; start is not automatic template generatio
|
|
|
140
142
|
|
|
141
143
|
Do not silently pick provider, model, effort or bypass. False preserves native permissions; it is not a read-only or sandbox guarantee. Use supported `models --provider NAME` discovery and retain exact model IDs.
|
|
142
144
|
|
|
143
|
-
Traditional minimal Markdown remains accepted by the original compiler. Its compatibility interpretation is **not a default for a new user plan**: provider defaults to Pi, bypass to false, and unset
|
|
145
|
+
Traditional minimal Markdown remains accepted by the original compiler. Its compatibility interpretation is **not a default for a new user plan**: provider defaults to Pi, bypass to false, and unset effort is left to the provider. Claude and Pi workers run only through their contract leaves, which require an exact model ID from `team-agent models --provider claude|pi`; an unset model is refused, never guessed. Team Agent does not infer a role model from TEAM defaults. Legacy explicit TEAM provider_effort remains supported for non-Pi roles; do not invent an effort for unspecified roles.
|
|
144
146
|
|
|
145
147
|
Supported role metadata:
|
|
146
148
|
- `agent_id` (legacy `name`) and `role` identify the seat.
|
|
@@ -155,7 +157,7 @@ A RoleOnly seat can be registered from its complete role file. A genuinely absen
|
|
|
155
157
|
|
|
156
158
|
```sh
|
|
157
159
|
team-agent start auditor --role-file /absolute/path/to/auditor.md
|
|
158
|
-
team-agent start analyst --provider pi --bypass false --prompt "Analyze the proposed change"
|
|
160
|
+
team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt "Analyze the proposed change"
|
|
159
161
|
```
|
|
160
162
|
|
|
161
163
|
An imported external role goes into this team's `agents/ID.md`; the original stays untouched. A file already at that destination is used in place. Equal file/CLI definitions are accepted; conflicting provider/model/effort/bypass/prompt/profile values refuse before writes/startup. Omitted model, effort, profile and role body remain unchanged.
|
|
@@ -18,6 +18,8 @@ team-agent agent restart reviewer --session preserve --workspace . --team projec
|
|
|
18
18
|
team-agent doctor --workspace . --json
|
|
19
19
|
team-agent inbox leader --workspace . --team project-team -n 3
|
|
20
20
|
team-agent models --provider NAME
|
|
21
|
+
team-agent models --provider claude
|
|
22
|
+
team-agent models --provider pi
|
|
21
23
|
team-agent profile init my-api --auth-mode compatible_api --workspace .
|
|
22
24
|
team-agent restart reviewer --workspace . --team project-team
|
|
23
25
|
team-agent route add pi -- --verbose
|
|
@@ -33,7 +35,7 @@ team-agent start --from ./my-team --workspace . --team project-team
|
|
|
33
35
|
team-agent start --from .team/current --workspace . --team project-team
|
|
34
36
|
team-agent start --help
|
|
35
37
|
team-agent start --workspace . --team project-team
|
|
36
|
-
team-agent start analyst --provider pi --bypass false --prompt "Analyze the proposed change"
|
|
38
|
+
team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt "Analyze the proposed change"
|
|
37
39
|
team-agent start auditor --role-file /absolute/path/to/auditor.md
|
|
38
40
|
team-agent start reviewer --effort high --bypass false
|
|
39
41
|
team-agent start reviewer --effort high --bypass false --workspace . --team project-team
|
|
@@ -140,7 +142,7 @@ Creation needs complete configuration; start is not automatic template generatio
|
|
|
140
142
|
|
|
141
143
|
Do not silently pick provider, model, effort or bypass. False preserves native permissions; it is not a read-only or sandbox guarantee. Use supported `models --provider NAME` discovery and retain exact model IDs.
|
|
142
144
|
|
|
143
|
-
Traditional minimal Markdown remains accepted by the original compiler. Its compatibility interpretation is **not a default for a new user plan**: provider defaults to Pi, bypass to false, and unset
|
|
145
|
+
Traditional minimal Markdown remains accepted by the original compiler. Its compatibility interpretation is **not a default for a new user plan**: provider defaults to Pi, bypass to false, and unset effort is left to the provider. Claude and Pi workers run only through their contract leaves, which require an exact model ID from `team-agent models --provider claude|pi`; an unset model is refused, never guessed. Team Agent does not infer a role model from TEAM defaults. Legacy explicit TEAM provider_effort remains supported for non-Pi roles; do not invent an effort for unspecified roles.
|
|
144
146
|
|
|
145
147
|
Supported role metadata:
|
|
146
148
|
- `agent_id` (legacy `name`) and `role` identify the seat.
|
|
@@ -155,7 +157,7 @@ A RoleOnly seat can be registered from its complete role file. A genuinely absen
|
|
|
155
157
|
|
|
156
158
|
```sh
|
|
157
159
|
team-agent start auditor --role-file /absolute/path/to/auditor.md
|
|
158
|
-
team-agent start analyst --provider pi --bypass false --prompt "Analyze the proposed change"
|
|
160
|
+
team-agent start analyst --provider pi --model openai-codex/gpt-6-luna --bypass false --prompt "Analyze the proposed change"
|
|
159
161
|
```
|
|
160
162
|
|
|
161
163
|
An imported external role goes into this team's `agents/ID.md`; the original stays untouched. A file already at that destination is used in place. Equal file/CLI definitions are accepted; conflicting provider/model/effort/bypass/prompt/profile values refuse before writes/startup. Omitted model, effort, profile and role body remain unchanged.
|