feat(loadtesting): agent guidance to cut redundant tool calls (prod) - #394
Merged
Conversation
Add capability guidance (and a product-summary note) that fixes the efficiency issues QA found on the agent surface: - report: prefer narrow slices (metrics / errorCategory / sinceIso / detail=per-txn) over refetching the full report; fetch full only when the user asks for the raw report. - getLoadTest / trends / run-history / active-runs: one call returns the whole result — reuse it within a task instead of re-fetching. - estimateLoadTestRunCost: always use it for cost/VU-hour math instead of computing by hand (misses ramp-up). - compareLoadTestRuns: use it for run deltas instead of diffing two reports. - getLoadTestRunStatus: a status question needs status only, not the report. - summary: state that delete and scheduling are not offered (point to the dashboard) so the agent stops searching to confirm the gap. Guidance-only; no endpoint or schema changes.
sarve-shreyas
merged commit Sep 7, 2026
5215889
into
browserstack:feat/capability-registry
1 check passed
This was referenced Sep 8, 2026
Closed
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follows up the prod capability index (#388, already merged) with guidance-only edits addressing the tool-call efficiency issues QA found on the agent surface (LTAI-270). No endpoint, schema, or
base_url/authchanges.What & why
QA's positive-path runs answered correctly but thrashed the generic discover→invoke layer — repeated identical invokes and over-discovery (e.g.
/report×4,/trends×8,/projects/{id}/runs×9). The endpoints already expose every slice/filter needed, so this steers the agent to use them and to reuse results within a task.metrics/errorCategory/slaOnly/sinceIso+untilIso/detail=per-txn+groupBy) instead of re-fetching the full report; full only when the user asks for the raw report.Scope
Guidance/
summarytext only — 20 capabilities unchanged in shape. The deeper lever (per-turn invoke dedup/caching in the generic host) is out of scope for the index and tracked separately.Regression companion: #389. Ref: LTAI-270.