feat(skills): native agentskills.io support#1
Closed
darkrishabh wants to merge 2 commits into
Closed
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
cursor
Bot
force-pushed
the
cursor/agent-skills-standard-6cb3
branch
from
bf427ee to
045dc91
Compare
cursor
Bot
changed the title
…op mode, HTML report
- Add structured SkillsEvent stream (suite/eval start/end) so consumers
can build rich UIs and reporters without re-parsing logs.
- Bundle a default consoleReporter with color-coded headers, prompt and
output snippets, per-assertion verdicts with evidence, and suite-level
pass-rate delta across with_skill / without_skill modes.
- Default workspace layout is now flat: <workspace>/<skill-slug>/...,
overwritten on each run (CI-friendly, easy to diff).
- Opt-in loop mode (loop: true) snapshots each run into
<workspace>/.history/iteration-N/<skill-slug>/... for skill-improvement
workflows, returning historyIteration in the result.
- Persist meta.json (skill metadata) and prompts.json (system, user, and
judge prompts actually sent) alongside the spec-mandated artifacts so
reports and debugging surface exactly what each model saw.
- Add a self-contained static HTML report (report.ts) at
<workspace>/report/index.html with overall stats, per-skill drill-down,
side-by-side with_skill / without_skill output, and a collapsible
judge prompt section.
- gradeOutputs now returns { grading, judgePrompt, judgeResponse } so the
judge prompt can be exposed in events and reports.
- evaluate-skills emits events, defaults to consoleReporter when no
onEvent/onLog is provided, generates the HTML report unless
report: false, and returns historyIteration + reportPath in
EvaluateSkillsResult.
- src/cli/skills.ts: update CLI JSON output to the new result shape.
darkrishabh
force-pushed
the
cursor/agent-skills-standard-6cb3
branch
from
045dc91 to
eb0ec5c
Compare
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.
Summary
@darkrishabh/bench-ai/skillsentrypoint:loadSkill,discoverSkills,runEval,gradeOutputs,writeRunArtifacts,buildBenchmark,ensureIterationDir, andevaluateSkills.@darkrishabh/bench-ai/providersfor provider imports and extend the Provider contract with optionalcapabilitiesandcompleteChatwhile preservingcomplete(prompt)compatibility.bench-ai skills <root>CLI support, spec-shaped artifact writing, baseline mode, README examples, a fixture skill, tests, and a minor version bump to1.1.0.Spec: https://agentskills.io/skill-creation/evaluating-skills
Testing
npm testnpm run type-checknpm run buildnode ./bench-ai.mjs skills --helpnode --input-type=module -e 'import { evaluateSkills } from "./dist/skills/index.js"; import { OpenAICompatibleProvider } from "./dist/providers/index.js"; console.log(typeof evaluateSkills, typeof OpenAICompatibleProvider);'