| Age | Commit message (Collapse) | Author |
|---|
|
A background subagents.lua workflow outlives the tool call that started
it, so settling the outer result unpinned a board whose agents were still
working. Track a per-tool-call workflow count and derive the pin from
both the tool being active and that count, so the board only returns to
transcript order once its last workflow finishes.
Also show the reasoning level alongside the model on agent cards.
|
|
|
|
Model-authored workflow Lua could reach string, table, math, and utf8 but
had no way to parse an agent's report, so any fan-out over discovered work
had to invent an ad-hoc line format and hand-roll string.gmatch. Hand the
guest panto.ext.json instead. decode returns nil plus a message rather
than raising: the sandbox deliberately has no pcall, so a raising decoder
would let one malformed agent line kill an entire workflow.
Teach the pattern in the injected guidance too: deterministic Lua owns the
loops, counts, ordering, and gates, agents supply discovery and judgment,
and a worker that feeds Lua ends its report with one parseable line. The
existing static-graph example stays; a discover-then-fan-out example joins
it.
|
|
subagents.lua now starts a workflow on its own coroutine and returns a
session-scoped id immediately, so fan-out continues while the primary keeps
working; completion wakes the primary, and later calls read immutable records
from subagents.workflows. Every ctx:agent takes a workflow-unique name so those
records are addressable. A session_start guidance message tells the primary when
to reach for run vs lua.
Children no longer inherit the primary's system context: a child starts from the
fixed child-role instruction plus its profile, and subagents.run/ctx:agent
accept an inline system_prompt instead of a profile. The now-redundant `agent`
form of subagents.models is gone.
Concurrency defaults to five and is configurable through [subagents]
max_concurrent in any layered config.toml; turn boundaries reap only settled
jobs so background workflows survive, while interrupt and session end cancel.
Config roots come from panto.ext.dirs.layers rather than a hand-rolled XDG
lookup, which picks up the base and git-ignored local layers for both agents/
and workflows/.
TOML workflows tighten up: the subagents.workflow tool takes a discovered name
only (inline `steps` duplicated subagents.lua at less power), an optional
top-level `output` array chooses the reported steps and their order instead of
the terminal set, a step with no workflow input gets no empty input heading, and
a workflow naming an undiscovered agent is rejected at discovery rather than
part-way through a run.
|
|
Record durable child-turn metadata, restore bounded progress cards during
startup replay, and keep settled cards attached to transcript entries. Add
workflow-local Lua profiles and expose discovered agents and workflows in the
session header.
|
|
The rock now owns all subagent policy on top of libpanto-lua's generic APIs:
children are ordinary panto.agent instances over rock-constructed stores,
started with agent:run_async and awaited by arming uv.new_poll on each job's
wake_fd from the tool handler's coroutine.
subagents/jobs.lua carries the session policy the host used to own: the
concurrency gate (4 running, FIFO queue, cancel-while-queued never starts),
the await contract (results in input order; "first" returns settled plus
remaining by identity), and settle-time shaping. subagents/spawn.lua seeds
new children (primary system context, child role, profile body with manifest
metadata), resolves model/reasoning through panto.ext.resolve_model, filters
subagents.* out of the inherited tool set via agent:set_tools, and reads
resume defaults back from stored message metadata. One-shot structured
workers are a null_store agent with a declaration-only output tool,
tool_choice forced, dispatch_tools=false. subagents/progress.lua renders
per-tool-entry cards through the component handle's invalidate seam;
turn_interrupt cancels live children, turn_end closes them.
Spec suite rewritten against fakes of the new surfaces (98 cases), including
gate/queue/cancel bounds, resume-default extraction, one-shot capture via
unresolved tool calls, tool filtering, and manifest seeding.
|
