Algis DumbrisandClaude Opus 4.6
ff5d0c49f4
Ships the MVP slice of spec 018 (dynamic agent spawning):
- 5 new SQLite migrations (021-025): goals + goal_tasks + agent_proposals
+ reputation_evidence + secrets + harness_runs.task_id. The legacy
`tasks` table (channel auctions) and `agent_trust` table (reactions
workflow) are left untouched — the new schema coexists.
- 4 new internal packages, fully tested:
- internal/goals: Goal struct + store + service, slug collision dedup,
backing-channel auto-create via ChannelCreator adapter
- internal/goaltasks: goal_tasks table with denormalized 16 KB
ancestry snapshots, single-statement optimistic-lock atomic claim,
recursive-CTE cost rollup, state machine, per-billing-code rollup
- internal/secrets: NaCl-secretbox encrypted blobs, user/agent/task
scope precedence, sanitized env injection, master-key bootstrap
- internal/trust additions: ConfigHash (deterministic SHA-256 of
model + prompt + tools + skills + mcp + subagents, sorted),
DelegationCap (tier + tool-scope + budget + depth enforcement),
append-only Ledger with exponential time-decay rolling score and
70%-of-parent child seeding. Existing trust package unchanged.
- Critical invariants under test:
- 50-goroutine concurrent claim race → exactly one winner per round
- ConfigHash stable under shuffled array inputs, sensitive to
capability changes
- DelegationCap full tier × tool-scope matrix
- Ledger time-decay + parent seed at 70 % ± 1 %
- Secret name sanitization, scope precedence, plaintext never
returned via MCP-equivalent paths
- internal/agents/types.go extended with dynamic-spawning columns
(config_hash, parent_agent_id, spawn_depth, system_prompt,
autonomy_tier, tool_scope_json, quarantined_at). Existing tests
still pass.
- cmd/docgardener: self-contained demo binary driving the end-to-end
flow. `docgardener run` creates a goal, builds a task tree with
denormalized ancestry, spawns 3 specialists (each going through
real delegation-cap validation and config-hash computation and
70 %-of-parent reputation seeding), claims tasks atomically, runs
them through the state machine, records reputation evidence.
`docgardener report` queries all of that back out and renders a
rich dark-mode HTML report (header, spend metrics, task tree,
spawned-agent cards with reputation bars, cost breakdown, artifacts,
timeline).
- examples/doc-gardener: start.sh / run_task.sh / report.sh / stop.sh
mirroring the cold-topic-explainer pattern. Launches an isolated
synapbus instance on port 18089, drives the demo, renders
report.html, cleans up. Full README documenting what's real vs
deferred, plus examples/README.md listing both examples.
- specs/018: tasks.md updated with MVP completion status; legacy tasks
naming collision noted.
Deferred (marked explicitly in example README):
- Real LLM-driven coordinator (needs MCP tool wiring + prompt
iteration)
- Real subprocess runs (needs reactor integration with task_id on
ExecRequest)
- Full MCP tool surface (contracts are written at
specs/018-dynamic-agent-spawning/contracts/mcp-tools.md)
- Svelte /goals UI (REST endpoints remain a follow-up)
- Full budget race + quarantine auto-trigger wiring
- Full resource-request → secrets fulfill reaction-workflow path
Cross-compiles clean for linux/amd64 and darwin/arm64 with no CGO
(SC-010). All new package tests pass (SC-004, SC-005, SC-007).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
SynapBus examples
Runnable demos of SynapBus features. Each example is self-contained under its own directory, launches an isolated synapbus instance on a distinct port, and cleans up after itself.
| Example | Feature | Real LLM? | Port |
|---|---|---|---|
cold-topic-explainer/ |
Reactive agent triggers + subprocess harness — three Gemini agents (decomposer → writer → critic) collaborate via DMs to produce a 3-paragraph explainer, with real LLM calls end-to-end. | ✅ yes (gemini CLI) |
18088 |
doc-gardener/ |
Dynamic agent spawning (spec 018) — a coordinator meta-agent decomposes a goal into a task tree, spawns specialists with config_hash-rooted trust + delegation-cap enforcement, runs them through the state machine, generates a rich HTML report. |
❌ v1 is synthetic (primitives demo); real LLM coordinator is a follow-up PR | 18089 |
Quick start
Pick an example, cd into it, and follow its README. In general:
cd examples/<name>
./start.sh # rebuild + launch an isolated synapbus instance
./run_task.sh # drive the demo flow
./report.sh # (where applicable) render an HTML report
./stop.sh # shut down
Both examples use the same layout for consistency:
examples/<name>/
├── start.sh # build & launch
├── run_task.sh # execute the demo flow
├── stop.sh # shut down
├── report.sh # (doc-gardener only) render HTML report
├── bin/
│ ├── synapbus # built from the current checkout
│ └── <helper> # example-specific driver binary
├── configs/ # per-agent JSON configs (harness_config, prompts, etc.)
├── data/ # isolated SQLite DB + attachment store + sockets
├── synapbus.log # server stdout+stderr
└── README.md # example-specific docs
What each example proves
- cold-topic-explainer proves that the SynapBus reactor + subprocess harness can drive a real multi-agent loop with three distinct LLMs, with depth and budget guards, OpenTelemetry tracing, and harness_runs accounting.
- doc-gardener proves that the dynamic-agent-spawning data primitives —
goals,goal_taskswith denormalized ancestry, atomic optimistic-lock claim,config_hash-keyed reputation ledger, delegation-cap enforcement, per-billing-code cost rollup — work end-to-end against real SQLite, and feed a rich HTML report.
The two examples are complementary: cold-topic-explainer exercises the runtime path (reactor → harness → LLM → DMs), doc-gardener exercises the work-tracking path (goals → tasks → trust → report). A future example will combine them into a full LLM-driven coordinator loop.
Global prereqs
- Go 1.25+
sqlite3,curl,jqon$PATH- A free TCP port per example (see table above)
- For
cold-topic-explaineronly:geminiCLI authenticated viagemini auth login
Troubleshooting
- Port already in use: set
SYNAPBUS_PORT=18090 ./start.sh(each example honors the env var). - Web UI is blank: rebuild the embedded Svelte SPA with
make webfrom the repo root once, then re-run./start.sh. - Stale binary: delete the example's
bin/directory and rerun./start.shto force a rebuild.