Skip to content

Actions: hookdeck/evals

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
50 workflow runs
50 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Refresh eval results Refresh eval results #15: Scheduled
4h 18m 11s main
main
4h 18m 11s
Report on the scenario criteria we wrote down and never checked Checks #35: Pull request #72 opened by leggetter
40s scenario-criteria-report
scenario-criteria-report
40s
Record what the published score counts, and what checks are for Checks #34: Pull request #71 opened by leggetter
34s record-scoring-definition
record-scoring-definition
34s
Let a scenario declare that one attempt is not enough for it Checks #33: Pull request #70 opened by leggetter
36s scenario-attempt-floor
scenario-attempt-floor
36s
Bring the plan's status, counts and costs back to what is true Checks #32: Pull request #69 opened by leggetter
36s plan-status-refresh
plan-status-refresh
36s
Score outpost-001 against the customer's endpoint, not one the agent picked Checks #31: Pull request #67 synchronize by leggetter
39s outpost-001-customer-endpoint
outpost-001-customer-endpoint
39s
Stop pnpm check from running scripts that no longer exist Checks #30: Pull request #68 opened by leggetter
43s fix-pnpm-check
fix-pnpm-check
43s
Score outpost-001 against the customer's endpoint, not one the agent picked Checks #29: Pull request #67 opened by leggetter
45s outpost-001-customer-endpoint
outpost-001-customer-endpoint
45s
Record what each published row was measured under Checks #28: Pull request #65 opened by leggetter
43s snapshot-provenance
snapshot-provenance
43s
Report how each arm reached the docs, and how often it got nothing Checks #27: Pull request #64 opened by leggetter
45s record-docs-reach
record-docs-reach
45s
Record what the do-not-ask instruction bought Checks #26: Pull request #63 opened by leggetter
45s record-ask-ab-result
record-ask-ab-result
45s
Tell agents not to ask clarifying questions Checks #25: Pull request #62 opened by leggetter
48s instruct-agents-not-to-ask
instruct-agents-not-to-ask
48s
Correct the weak-model skills delta to −2 Checks #24: Pull request #59 opened by leggetter
41s correct-skills-figure
correct-skills-figure
41s
triage: flag an agent that stopped to ask a question Checks #23: Pull request #58 synchronize by leggetter
42s triage-asked-and-stopped
triage-asked-and-stopped
42s
triage: flag an agent that stopped to ask a question Checks #22: Pull request #58 opened by leggetter
38s triage-asked-and-stopped
triage-asked-and-stopped
38s
Persist the Outpost config baseline, and let a run be told not to publish Checks #21: Pull request #56 opened by leggetter
39s persist-config-baseline
persist-config-baseline
39s
Correctness sweep: dead processes, wildcards, an inert field, and --resume Checks #20: Pull request #55 opened by leggetter
46s correctness-sweep
correctness-sweep
46s
Skip a scenario whose seed credentials are missing, rather than failing it Checks #19: Pull request #54 opened by leggetter
45s skip-without-seed-credentials
skip-without-seed-credentials
45s
Publish the Outpost snapshot: 104/114 Checks #18: Pull request #53 opened by leggetter
46s release-0.3.0
release-0.3.0
46s
outpost-004: stop shipping AWS's documented example credentials Checks #17: Pull request #52 opened by leggetter
45s outpost-004-credentials
outpost-004-credentials
45s
Correct three things the Outpost runs showed were wrong Checks #16: Pull request #44 synchronize by leggetter
43s outpost-scorer-corrections
outpost-scorer-corrections
43s
Correct three things the Outpost runs showed were wrong Checks #15: Pull request #44 synchronize by leggetter
18s outpost-scorer-corrections
outpost-scorer-corrections
18s
Correct three things the Outpost runs showed were wrong Checks #14: Pull request #44 synchronize by leggetter
39s outpost-scorer-corrections
outpost-scorer-corrections
39s
Refresh eval results Refresh eval results #14: Scheduled
3h 0m 40s main
main
3h 0m 40s
Correct three things the Outpost runs showed were wrong Checks #13: Pull request #44 synchronize by leggetter
33s outpost-scorer-corrections
outpost-scorer-corrections
33s