feat(agent-interface): harness selector-support capability (gtm#396)#10
Merged
Merged
Conversation
Add harnessHonorsModel / harnessHonorsEffort / harnessHonorsSelectors — whether a harness's runner actually honors the per-turn model and reasoning-effort pickers. Grounded in the cli-bridge adapter audit (amp drops both; openclaw/nanoclaw drop the model; factory-droids/hermes/nanoclaw drop the effort); everything else honors both. Lets a chat picker trim or mark harnesses up front so a user's model/effort choice is never silently ignored. Distinct from reasoningEffortsFor (which levels a harness can express). Follows the existing table-per-capability pattern; aliases canonicalized.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Adds a harness selector-support capability to the single source of truth, for Problem 3 of
gtm-agent#396 (harnesses that ignore the chat pickers):
harnessHonorsModel(harness)— does the runner honor a per-turn model override?harnessHonorsEffort(harness)— does it honor a reasoning-effort override?harnessHonorsSelectors(harness)— the AND of both (model + effort pickers are live).Grounded in the cli-bridge adapter audit, not a guess. Most harnesses honor both, so only the
exceptions are encoded (same table-per-capability style as
harnessProviderLock/harnessReasoningCeiling):Aliases are canonicalized before keying (e.g.
claude→claude-code).Why
This is distinct from
reasoningEffortsFor(which levels a harness can express). A chat pickerneeds to know, up front, whether to even offer a harness — otherwise a user's model/effort choice is
silently dropped (e.g. picking a model on
ampdoes nothing). The sandbox-ui chat picker will consumethese to trim or mark such harnesses.
Verification
tsc --noEmitclean,buildclean, new symbols present indist.per-turn selector supportdescribe block).Release
@tangle-network/agent-interfaceminor (0.15.0 → 0.16.0).