From 8cb291ad7cdb317ff80947278ee055b1a4925b41 Mon Sep 17 00:00:00 2001 From: Claude Sonnet 5 Date: Sat, 4 Jul 2026 05:05:36 +0000 Subject: feat(story,role): add retro ceremony -- closes the self-improvement loop (Phase 8) The final mechanism the versioned role-config model (Phase 5) was built for: when a story reaches DONE, StoryOrchestrator spawns a retro-role task that reflects on the story's full history and proposes draft role_configs versions for a human to review and activate via the existing (unchanged) POST /api/roles/{role}/activate. - AgentChannel gains a 6th method, ProposeRoleConfig(ctx, role.RoleConfig) (version, err), following ProposeEpic's precedent (Phase 7c): a structured tool call, not summary-parsing. storeChannel.ProposeRoleConfig calls the same Store.CreateRoleConfig the human-facing POST /api/roles/{role}/versions endpoint already uses (proposed_by: "retro"), landing a new draft row without touching whatever's currently active. Wired through both transports exactly like ProposeEpic: internal/agentloop/tools.go (native loop) and internal/executor/agentmcp.go (MCP). - StoryOrchestrator.Tick now routes a story at status DONE to a new processRetro stage instead of processStory -- a sibling stage, not a continuation, since the Builder->Evaluators->Arbitration chain is long settled by then. processRetro only *reads* that settled pipeline (read-only findEvaluators/findArbitration counterparts to ensureEvaluators/ensureArbitration -- it never spawns/mutates Builder-pipeline tasks) to locate the Arbitration task the retro task depends on, then spawns (idempotently -- checks for an existing retro-role dependent first) one retro-role task with instructions assembled from the story's spec/acceptance-criteria, full task tree, per- task cost/escalation history, active role_configs per role encountered, and the story's own event stream (evaluator verdicts, arbitration decision). - event.KindRetroCaptured (attached to the story's ID, matching KindEvalVerdict/KindArbitrationDecided's convention) fires once the retro task completes (auto-accepted like every other pipeline task), aggregating every event.KindRoleConfigProposed the retro task recorded (one per propose_role_config call) into {task_id, proposals: [{role, version}], summary} -- the summary is the "capturing lessons" half of this ceremony, the proposals are the versioned-config half. - Human activation is completely untouched: drafts land through the identical CreateRoleConfig/config_json path Phase 5's endpoints already handle, confirmed via existing role-endpoint tests passing unmodified. go build/vet/test -race -count=1 all pass, full suite (20 packages) -- one run hit a known, pre-existing, intermittent flake under full-suite load (unrelated to this phase's files) that did not reproduce on two immediate reruns, both in isolation and full-suite. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_01V1moSNCJRcP6kykA4tyUSs --- internal/agentloop/tools.go | 63 ++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 56 insertions(+), 7 deletions(-) (limited to 'internal/agentloop') diff --git a/internal/agentloop/tools.go b/internal/agentloop/tools.go index d803c23..ad1aecf 100644 --- a/internal/agentloop/tools.go +++ b/internal/agentloop/tools.go @@ -8,16 +8,17 @@ import ( "github.com/thepeterstone/claudomator/internal/agentchannel" "github.com/thepeterstone/claudomator/internal/provider" + "github.com/thepeterstone/claudomator/internal/role" "github.com/thepeterstone/claudomator/internal/sandbox" ) -// agentToolSpecs returns the nine tools available to the loop: the five -// agent back-channel tools (mirroring the MCP tools ContainerRunner exposes; -// propose_epic was added in Phase 7c) plus the four sandbox tools, as -// provider-neutral ToolSpecs. The original four (ask_user/report_summary/ -// spawn_subtask/record_progress) plus the sandbox tools were ported verbatim -// (same names/descriptions/JSON schemas) from the former -// executor.agentToolDefs (internal/executor/localtools.go). +// agentToolSpecs returns the tools available to the loop: the agent +// back-channel tools (mirroring the MCP tools ContainerRunner exposes; +// propose_epic was added in Phase 7c, propose_role_config in Phase 8) plus +// the four sandbox tools, as provider-neutral ToolSpecs. The original four +// (ask_user/report_summary/spawn_subtask/record_progress) plus the sandbox +// tools were ported verbatim (same names/descriptions/JSON schemas) from the +// former executor.agentToolDefs (internal/executor/localtools.go). func agentToolSpecs() []provider.ToolSpec { strProp := func(desc string) map[string]any { return map[string]any{"type": "string", "description": desc} @@ -81,6 +82,45 @@ func agentToolSpecs() []provider.ToolSpec { "required": []string{"name", "story_ids"}, }, }, + { + Name: "propose_role_config", + Description: "Propose a new draft configuration version for a role, after reflecting on what happened (e.g. during a story retro). Creates a new draft role_configs row for a human to review and activate -- it never changes what is currently active.", + ParametersJSONSchema: map[string]any{ + "type": "object", + "properties": map[string]any{ + "role": strProp("the role name this config applies to (e.g. an existing role like builder, or a new one)"), + "system_prompt": strProp("system prompt appended for tasks dispatched through this role"), + "tools": map[string]any{"type": "array", "items": map[string]any{"type": "string"}, "description": "optional tool allowlist for this role"}, + "sandbox_kind": strProp("optional sandbox kind for this role"), + "default_budget_usd": map[string]any{"type": "number", "description": "optional estimated budget in USD, used when the scheduler considers escalating this role's tasks"}, + "escalation_ladder": map[string]any{ + "type": "array", + "description": "ordered list of escalation tiers", + "items": map[string]any{ + "type": "object", + "properties": map[string]any{ + "candidates": map[string]any{ + "type": "array", + "items": map[string]any{ + "type": "object", + "properties": map[string]any{ + "provider": strProp("executor runner key, e.g. anthropic, google, groq, openrouter, openai, local"), + "model": strProp("model name"), + }, + "required": []string{"provider", "model"}, + }, + "description": "candidate provider/model rungs for this tier", + }, + "selection_mode": strProp("round_robin (default) or single"), + "max_retries": map[string]any{"type": "integer", "description": "attempts allowed at this tier before escalating to the next"}, + }, + "required": []string{"candidates"}, + }, + }, + }, + "required": []string{"role"}, + }, + }, { Name: "read_file", Description: "Read the contents of a file in the sandbox working directory.", @@ -211,6 +251,15 @@ func (l *Loop) dispatchTool(ctx context.Context, name, argsJSON string) (result } return "Proposed epic " + id, false, nil + case "propose_role_config": + var a role.RoleConfig + _ = json.Unmarshal([]byte(argsJSON), &a) + version, prErr := l.Channel.ProposeRoleConfig(ctx, a) + if prErr != nil { + return "", false, prErr + } + return fmt.Sprintf("Proposed role config %s v%d (draft)", a.Role, version), false, nil + case "read_file": var a struct { Path string `json:"path"` -- cgit v1.2.3