Directories
¶
| Path | Synopsis |
|---|---|
|
Package approvalnotify wires the permission gate's unwatched-prompt signal to the operator's alert registry (#647).
|
Package approvalnotify wires the permission gate's unwatched-prompt signal to the operator's alert registry (#647). |
|
Package approvereval evaluates auto mode's approver against a labelled corpus (#1175 phase 5, docs/auto-mode-design.md "Evaluation").
|
Package approvereval evaluates auto mode's approver against a labelled corpus (#1175 phase 5, docs/auto-mode-design.md "Evaluation"). |
|
Package attachclient is the shared client for attach-mode endpoints, used by both `core-agent attach`/`ls` (in cmd/core-agent) and `core-agent-tui` (in cmd/core-agent-tui).
|
Package attachclient is the shared client for attach-mode endpoints, used by both `core-agent attach`/`ls` (in cmd/core-agent) and `core-agent-tui` (in cmd/core-agent-tui). |
|
Package coretuievent projects ADK session events into the shapes go-steer/core-tui renders.
|
Package coretuievent projects ADK session events into the shapes go-steer/core-tui renders. |
|
Package coretuiremote adapts a remote core-agent (reached over the pkg/attach HTTP+SSE protocol via internal/attachclient) into a go-steer/core-tui Agent.
|
Package coretuiremote adapts a remote core-agent (reached over the pkg/attach HTTP+SSE protocol via internal/attachclient) into a go-steer/core-tui Agent. |
|
Package effects implements the recorded-effect outbox (docs/durable-write-gate-design.md): the runtime guard that makes mutating-tool re-execution ambiguity visible and blocking instead of silent.
|
Package effects implements the recorded-effect outbox (docs/durable-write-gate-design.md): the runtime guard that makes mutating-tool re-execution ambiguity visible and blocking instead of silent. |
|
Package evals is the behavioural eval skeleton (#652): one case, run against a real provider, graded by a deterministic verifier.
|
Package evals is the behavioural eval skeleton (#652): one case, run against a real provider, graded by a deterministic verifier. |
|
cmd/evals
command
Command evals runs the parity and differentiator eval suite (#1403).
|
Command evals runs the parity and differentiator eval suite (#1403). |
|
cmd/outcome
command
Command outcome runs the O tier: the outcome evals (#1403, docs/mast-fold-design.md §3.8).
|
Command outcome runs the O tier: the outcome evals (#1403, docs/mast-fold-design.md §3.8). |
|
differentiators
Package differentiators holds the v0.3 eval scenarios that the upstream LangChain SRE harness structurally cannot express (docs/v0.3-plan.md W0.3).
|
Package differentiators holds the v0.3 eval scenarios that the upstream LangChain SRE harness structurally cannot express (docs/v0.3-plan.md W0.3). |
|
evalcompose
Package evalcompose provides the runtime wiring helpers used by the judge and outcome eval tiers: model construction, per-specialist budget metering, fan-out roster assembly, and the in-turn loop guard.
|
Package evalcompose provides the runtime wiring helpers used by the judge and outcome eval tiers: model construction, per-specialist budget metering, fan-out roster assembly, and the in-turn loop guard. |
|
harness
Package harness is the runnable form of the parity and differentiator eval suite (#1403): the thing dev/tools/evals invokes and CI (dev/ci/presubmits/evals) gates on.
|
Package harness is the runnable form of the parity and differentiator eval suite (#1403): the thing dev/tools/evals invokes and CI (dev/ci/presubmits/evals) gates on. |
|
judge
Package judge is the metered tier of the v0.3 parity eval suite (docs/v0.3-plan.md W0.5): the 31-scenario corpus scored against a real model, which is the only tier that produces a LangChain-comparable number.
|
Package judge is the metered tier of the v0.3 parity eval suite (docs/v0.3-plan.md W0.5): the 31-scenario corpus scored against a real model, which is the only tier that produces a LangChain-comparable number. |
|
outcome
Package outcome is the O tier's corpus: the fixture-role catalog, the case schema, and the loader that refuses a corpus which cannot measure what it claims to.
|
Package outcome is the O tier's corpus: the fixture-role catalog, the case schema, and the loader that refuses a corpus which cannot measure what it claims to. |
|
Package imagepin discovers where this repo pins a container image.
|
Package imagepin discovers where this repo pins a container image. |
|
Package kvwindow is the shape in which a model server's KV-cache activity is reported, and the fold that sums windows for the judged eval board (internal/evals/harness).
|
Package kvwindow is the shape in which a model server's KV-cache activity is reported, and the fold that sums windows for the judged eval board (internal/evals/harness). |
|
Package modelretry waits out a provider that says "not now".
|
Package modelretry waits out a provider that says "not now". |
|
Package modeltext reads the caller-facing text out of a model response part (#370).
|
Package modeltext reads the caller-facing text out of a model response part (#370). |
|
Package monitor parses the structured transition records returned by a monitor.collect tool and defines the argument contract for monitor.ack.
|
Package monitor parses the structured transition records returned by a monitor.collect tool and defines the argument contract for monitor.ack. |
|
Package notify implements the switchboard /v1/messages HTTP client used by monitor.notify to post, replace, and thread-append monitoring digests.
|
Package notify implements the switchboard /v1/messages HTTP client used by monitor.notify to post, replace, and thread-append monitoring digests. |
|
Package subagentlog reads one subagent's inner turns back out of the parent session's event log.
|
Package subagentlog reads one subagent's inner turns back out of the parent session's event log. |
|
Package testutil provides shared test helpers for Cogo, most notably FakeModel — a deterministic implementation of google.golang.org/adk/v2/model.LLM that lets us drive end-to-end agent tests without burning real tokens.
|
Package testutil provides shared test helpers for Cogo, most notably FakeModel — a deterministic implementation of google.golang.org/adk/v2/model.LLM that lets us drive end-to-end agent tests without burning real tokens. |
|
Package version centralizes build-identity reporting for both cmd/core-agent and cmd/core-agent-tui.
|
Package version centralizes build-identity reporting for both cmd/core-agent and cmd/core-agent-tui. |
|
Package webui embeds a pinned release of the mast-web operator UI (github.com/go-steer/mast-web) so the agent binary can serve it at /ui/* on its attach listener when the --ui flag is set.
|
Package webui embeds a pinned release of the mast-web operator UI (github.com/go-steer/mast-web) so the agent binary can serve it at /ui/* on its attach listener when the --ui flag is set. |
Click to show internal directories.
Click to hide internal directories.