cowork-flow 0.0.24 → 0.0.25

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.
Files changed (4) hide show
  1. package/package.json +1 -1
  2. package/template/.agents/skills/party-mode/SKILL.md +41 -10
  3. package/template/.claude/skills/party-mode/SKILL.md +41 -10
  4. package/template/.cowork-flow/.version +1 -1
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "cowork-flow",
3
- "version": "0.0.24",
3
+ "version": "0.0.25",
4
4
  "description": "CLI for installing and maintaining cowork-flow collaboration templates.",
5
5
  "type": "module",
6
6
  "bin": {
package/template/.agents/skills/party-mode/SKILL.md CHANGED
@@ -20,7 +20,7 @@ Use this skill as a manual advisory roundtable. Coordinate true child agents, no
20
20
  Built-in defaults:
21
21
 
22
22
  - `max_agents=3`
23
- - `max_rounds=3`
23
+ - `max_rounds=5`
24
24
 
25
25
  These defaults are configurable values, not hard-coded constants. Effective config order:
26
26
 
@@ -35,7 +35,7 @@ Configurable fields:
35
35
  - agent roster or review lenses
36
36
  - report enabled
37
37
  - report path
38
- - whether round 2 or round 3 is allowed
38
+ - round phase policy, including when challenge may continue and when convergence must begin
39
39
 
40
40
  Safety gates:
41
41
 
@@ -47,10 +47,14 @@ Safety gates:
47
47
  ## Round Model
48
48
 
49
49
  1. Frame the question, decision needed, scope, and evidence packet.
50
- 2. Round 1 uses fresh child contexts. Send each child the same packet and one lens. Child agents cannot see each other.
51
- 3. Synthesize only evidence-backed positions, disagreements, rejected options, and acceptance signals.
52
- 4. Continue only when a continue condition is met. Send narrow follow-up prompts to the smallest useful set of children.
53
- 5. Stop when any stop condition is met. Close all live children through the Host Adapter close primitive.
50
+ 2. Select the smallest useful agent roster or review lenses within effective `max_agents`. Record why each selected voice is useful and why omitted voices are not needed.
51
+ 3. Round 1 uses fresh child contexts. Send each child the same packet and one lens. Child agents cannot see each other.
52
+ 4. Synthesize evidence-backed positions into a compact claim table: `claim_id`, owner, claim, evidence, counterclaim, evidence gap, and decision impact.
53
+ 5. Continue only when a continue condition is met. Send narrow follow-up prompts to the smallest useful set of children, and bind each prompt to one `claim_id`.
54
+ 6. Follow-up rounds should prefer the same live child that produced or challenged the relevant position. Send only the target claim, counterclaim, evidence gap, and ask the child to `agree`, `reject`, or `revise`.
55
+ 7. For Challenge rounds, the default stance is scrutiny. Test the opposing claim first; choose `agree` only after naming the evidence that compels agreement.
56
+ 8. Spawn an extra child only when the effective roster or lens config allows it, a live child failed, or the user approves expansion.
57
+ 9. Stop when any stop condition is met and no dispatched child is still running. Then close live children through the Host Adapter close primitive.
54
58
 
55
59
  Round intent:
56
60
 
@@ -61,10 +65,10 @@ Round intent:
61
65
  Default mapping:
62
66
 
63
67
  - Round 1 = Opening.
64
- - Round 2 = Challenge.
65
- - Round 3+ = Convergence.
68
+ - Round 2+ = Challenge while continue conditions still expose material disagreement, material risk, missing evidence, or untestable acceptance criteria.
69
+ - Convergence begins when the coordinator can write one recommended direction and no material challenge condition remains.
66
70
 
67
- Only the user can restart exploration after convergence begins.
71
+ After convergence begins, do not reopen exploration unless the user approves or new concrete evidence appears.
68
72
 
69
73
  ## Continue Conditions
70
74
 
@@ -86,6 +90,13 @@ Stop when any condition holds:
86
90
  - Effective `max_rounds` is reached.
87
91
  - Output fails schema, and one repair prompt does not fix it.
88
92
 
93
+ ## Waiting Children
94
+
95
+ - A wait timeout is not a child timeout.
96
+ - If a child remains live after a wait timeout, wait again or report it in `pending_children`.
97
+ - Do not close, cancel, omit, or synthesize around a live child unless the user explicitly asks to close it.
98
+ - When the discussion has converged or the user ends it, close live children after their final output is recorded.
99
+
89
100
  ## Child Output Schema
90
101
 
91
102
  Each child must return these core fields. Extra fields are allowed only after them.
@@ -100,6 +111,18 @@ acceptance_signal:
100
111
  what_would_change_my_mind:
101
112
  ```
102
113
 
114
+ Follow-up rounds may add these fields after the core fields:
115
+
116
+ ```text
117
+ claim_id:
118
+ responding_to:
119
+ opposing_claim:
120
+ position_delta:
121
+ evidence_delta:
122
+ still_disagree:
123
+ ```
124
+
125
+ Use `position_delta` to say whether the child maintained, narrowed, or changed its position.
103
126
  Reject unsupported opinion. Evidence should name files, commands, rules, observed behavior, user scenarios, or concrete assumptions.
104
127
 
105
128
  ## Coordinator Output Schema
@@ -107,6 +130,13 @@ Reject unsupported opinion. Evidence should name files, commands, rules, observe
107
130
  Final synthesis must return these core fields. Extra fields are allowed only after them.
108
131
 
109
132
  ```text
133
+ effective_max_agents:
134
+ effective_max_rounds:
135
+ rounds_used:
136
+ selected_agents:
137
+ claim_table:
138
+ agent_turns:
139
+ pending_children:
110
140
  consensus:
111
141
  disagreements:
112
142
  evidence:
@@ -114,7 +144,8 @@ decision:
114
144
  rejected_options:
115
145
  acceptance_criteria:
116
146
  open_questions:
147
+ early_stop_reason:
117
148
  stop_reason:
118
149
  ```
119
150
 
120
- Keep the final decision traceable to child evidence. Do not count votes as validation.
151
+ `selected_agents` must include the selected agent or lens names and selection reasons. `agent_turns` should preserve a compact transcript with round, agent or lens, `claim_id`, position, and `position_delta`. Keep the final decision traceable to child evidence. Do not count votes as validation.
package/template/.claude/skills/party-mode/SKILL.md CHANGED
@@ -20,7 +20,7 @@ Use this skill as a manual advisory roundtable. Coordinate true child agents, no
20
20
  Built-in defaults:
21
21
 
22
22
  - `max_agents=3`
23
- - `max_rounds=3`
23
+ - `max_rounds=5`
24
24
 
25
25
  These defaults are configurable values, not hard-coded constants. Effective config order:
26
26
 
@@ -35,7 +35,7 @@ Configurable fields:
35
35
  - agent roster or review lenses
36
36
  - report enabled
37
37
  - report path
38
- - whether round 2 or round 3 is allowed
38
+ - round phase policy, including when challenge may continue and when convergence must begin
39
39
 
40
40
  Safety gates:
41
41
 
@@ -47,10 +47,14 @@ Safety gates:
47
47
  ## Round Model
48
48
 
49
49
  1. Frame the question, decision needed, scope, and evidence packet.
50
- 2. Round 1 uses fresh child contexts. Send each child the same packet and one lens. Child agents cannot see each other.
51
- 3. Synthesize only evidence-backed positions, disagreements, rejected options, and acceptance signals.
52
- 4. Continue only when a continue condition is met. Send narrow follow-up prompts to the smallest useful set of children.
53
- 5. Stop when any stop condition is met. Close all live children through the Host Adapter close primitive.
50
+ 2. Select the smallest useful agent roster or review lenses within effective `max_agents`. Record why each selected voice is useful and why omitted voices are not needed.
51
+ 3. Round 1 uses fresh child contexts. Send each child the same packet and one lens. Child agents cannot see each other.
52
+ 4. Synthesize evidence-backed positions into a compact claim table: `claim_id`, owner, claim, evidence, counterclaim, evidence gap, and decision impact.
53
+ 5. Continue only when a continue condition is met. Send narrow follow-up prompts to the smallest useful set of children, and bind each prompt to one `claim_id`.
54
+ 6. Follow-up rounds should prefer the same live child that produced or challenged the relevant position. Send only the target claim, counterclaim, evidence gap, and ask the child to `agree`, `reject`, or `revise`.
55
+ 7. For Challenge rounds, the default stance is scrutiny. Test the opposing claim first; choose `agree` only after naming the evidence that compels agreement.
56
+ 8. Spawn an extra child only when the effective roster or lens config allows it, a live child failed, or the user approves expansion.
57
+ 9. Stop when any stop condition is met and no dispatched child is still running. Then close live children through the Host Adapter close primitive.
54
58
 
55
59
  Round intent:
56
60
 
@@ -61,10 +65,10 @@ Round intent:
61
65
  Default mapping:
62
66
 
63
67
  - Round 1 = Opening.
64
- - Round 2 = Challenge.
65
- - Round 3+ = Convergence.
68
+ - Round 2+ = Challenge while continue conditions still expose material disagreement, material risk, missing evidence, or untestable acceptance criteria.
69
+ - Convergence begins when the coordinator can write one recommended direction and no material challenge condition remains.
66
70
 
67
- Only the user can restart exploration after convergence begins.
71
+ After convergence begins, do not reopen exploration unless the user approves or new concrete evidence appears.
68
72
 
69
73
  ## Continue Conditions
70
74
 
@@ -86,6 +90,13 @@ Stop when any condition holds:
86
90
  - Effective `max_rounds` is reached.
87
91
  - Output fails schema, and one repair prompt does not fix it.
88
92
 
93
+ ## Waiting Children
94
+
95
+ - A wait timeout is not a child timeout.
96
+ - If a child remains live after a wait timeout, wait again or report it in `pending_children`.
97
+ - Do not close, cancel, omit, or synthesize around a live child unless the user explicitly asks to close it.
98
+ - When the discussion has converged or the user ends it, close live children after their final output is recorded.
99
+
89
100
  ## Child Output Schema
90
101
 
91
102
  Each child must return these core fields. Extra fields are allowed only after them.
@@ -100,6 +111,18 @@ acceptance_signal:
100
111
  what_would_change_my_mind:
101
112
  ```
102
113
 
114
+ Follow-up rounds may add these fields after the core fields:
115
+
116
+ ```text
117
+ claim_id:
118
+ responding_to:
119
+ opposing_claim:
120
+ position_delta:
121
+ evidence_delta:
122
+ still_disagree:
123
+ ```
124
+
125
+ Use `position_delta` to say whether the child maintained, narrowed, or changed its position.
103
126
  Reject unsupported opinion. Evidence should name files, commands, rules, observed behavior, user scenarios, or concrete assumptions.
104
127
 
105
128
  ## Coordinator Output Schema
@@ -107,6 +130,13 @@ Reject unsupported opinion. Evidence should name files, commands, rules, observe
107
130
  Final synthesis must return these core fields. Extra fields are allowed only after them.
108
131
 
109
132
  ```text
133
+ effective_max_agents:
134
+ effective_max_rounds:
135
+ rounds_used:
136
+ selected_agents:
137
+ claim_table:
138
+ agent_turns:
139
+ pending_children:
110
140
  consensus:
111
141
  disagreements:
112
142
  evidence:
@@ -114,7 +144,8 @@ decision:
114
144
  rejected_options:
115
145
  acceptance_criteria:
116
146
  open_questions:
147
+ early_stop_reason:
117
148
  stop_reason:
118
149
  ```
119
150
 
120
- Keep the final decision traceable to child evidence. Do not count votes as validation.
151
+ `selected_agents` must include the selected agent or lens names and selection reasons. `agent_turns` should preserve a compact transcript with round, agent or lens, `claim_id`, position, and `position_delta`. Keep the final decision traceable to child evidence. Do not count votes as validation.
package/template/.cowork-flow/.version CHANGED
@@ -1 +1 @@
1
- 0.0.24
1
+ 0.0.25