Repository navigation
feat(instructions): make plain language a shared HVE default - #3241
Open
Bill Berry (WilliamBerryiii) wants to merge 2 commits into
Open
Bill Berry (WilliamBerryiii) wants to merge 2 commits into
Bill Berry (WilliamBerryiii) wants to merge 2 commits into
Conversation
- add direct reminders and shared guidance across agents and skills - preserve technical formats and add six advisory writing checks - refresh package metadata and reference documentation 🗣️ - Generated by Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3241 +/- ##
==========================================
+ Coverage 81.90% 82.01% +0.11%
==========================================
Files 125 126 +1
Lines 17625 17775 +150
Branches 63 63
==========================================
+ Hits 14436 14579 +143
- Misses 3178 3185 +7
Partials 11 11
Flags with carried forward coverage won't be shown. Click here to find out more. 🚀 New features to boost your workflow:
|
Contributor
Eval Execution❌ Status: Failed — 2 spec(s) block merge
|
- add six source-backed advisory skill conformance cases - document canonical changed-artifact planning preflight 🧪 - Generated by Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Pull Request
Description
Make plain language a shared HVE default for questions, explanations, progress updates, summaries, and generated prose without changing exact technical formats.
The CI follow-up adds six source-backed conformance cases for
performance-slo-plannerandtelemetry-foundations, whose missing coveragecaused Agent Eval Plan to fail. The planner and its fail-closed gate are
unchanged. The eval guide now includes deterministic coverage and canonical
planning checks before model execution.
Related Issue(s)
Closes #3237
Type of Change
Select all that apply:
Code & Documentation:
Infrastructure & Configuration:
AI Artifacts:
hve-builderand addressed all actionable findings.github/instructions/*.instructions.md).github/prompts/*.prompt.md).github/agents/*.agent.md).github/skills/*/SKILL.md).github/hooks/*/*.json)evals/)Other:
.ps1,.sh,.py)Sample Prompts (for AI Artifact Contributions)
User Request:
"Explain why retrying this import will not create duplicate rows."
Execution Flow:
The selected agent or skill carries a direct writing reminder and obtains the shared Plain Language instruction. It follows its existing workflow and preserves required technical formats.
Output Artifacts:
A direct explanation, for example:
Success Indicators:
The reader understands the result without knowing internal workflow labels. Necessary technical names and machine-only formats remain unchanged. Actual request inclusion and response quality must be verified separately.
For detailed contribution requirements, see:
Testing
HVE Builder authoring, validation, and artifact-quality review completed with required findings resolved. The final candidate preserves all 151 original entry bodies apart from the approved reminders and existing date updates.
npm cinpm run lint:frontmatternpm run lint:mdnpm run lint:tablesnpm run lint:yamlnpm run lint:artifact-portabilitynpm run lint:cold-startnpm run validate:skillsnpm run lint:modelsnpm run lint:asset-docsnpm run lint:extension-artifact-namingnpm run lint:linksnpm run plugin:syncnpm run plugin:validatenpm run docs:generatenpm run docs:generate:checknpm run extension:preparenpm run lint:eval-grader-lineagepwsh -NoProfile -File scripts\evals\Test-EvalSpec.ps1 -Root evals\behavior-conformance -SkipAgentCoverage -OutputPath logs\plain-language-eval-spec-final.jsonpwsh -NoProfile -File scripts\evals\Test-VallyTestSafety.ps1 -OutputPath logs\plain-language-eval-safety-final.jsonpwsh -NoProfile -File scripts\evals\Build-AgentBehaviorSpec.ps1 -WhatIfpwsh -NoProfile -File scripts\evals\Test-EvalSpecText.ps1.\node_modules\.bin\vally.cmd lint --eval-spec evals\behavior-conformance\instructions.eval.yaml.\node_modules\.bin\cspell.cmd stdin --config .cspell.json --no-progress, with added/new-document text piped innpm run spell-checknpm run lint:md-links -- -ChangedFilesOnlynpm run lint:md-links -- -ChangedFilesOnly -ExternalLinksAsWarningsgit diff --check origin/main...HEADAdditional local checks verified six novel normalized prompt hashes, preserved all 80 existing instruction stimuli/grader identities, and exercised positive/negative/prompt-echo cases for the three new regex graders. A temporary extension stage matched all 85 shipped skill entries and resolved all 60 shipped agent policy imports; it was removed without creating a VSIX.
Model-backed evaluations, semantic judge calibration, native installed-host observations, hosted CI, the full local aggregate, and Docusaurus component/browser suites were not run. These static and source checks do not establish measured reduction in jargon across models.
CI repair validation
The original hosted failure was reproduced locally from its immutable CI
comparison and artifact manifest: exactly two skills had no covering specs.
The repaired tree reports 149 covered artifacts, zero missing, and five
intentional repository-only skips.
pwsh -NoProfile -File scripts\evals\Get-EvalChangeSet.ps1 -BaseRef ee3df3231ebf07e6f0c4b3ff1dd3a8f1f08e0134 -HeadRef HEAD -OutFile logs\repair-committed-change-set.jsonee11f7b1epwsh -NoProfile -File scripts\evals\Get-ChangedAIArtifact.ps1 -ChangeSetPath logs\repair-committed-change-set.json -OutFile logs\repair-committed-artifacts.jsonpwsh -NoProfile -File scripts\evals\Get-ChangedSpecStimulus.ps1 -ChangeSetPath logs\repair-committed-change-set.json -OutFile logs\repair-committed-stimuli.jsonpwsh -NoProfile -File scripts\evals\Test-StimulusPresence.ps1 -ManifestPath logs\repair-committed-artifacts.json -EvalRoot evals -OutFile logs\repair-committed-presence.json -FailOnSpecErrorpwsh -NoProfile -File scripts\evals\New-AgentEvalPlan.ps1 -ManifestPath logs\repair-committed-artifacts.json -ChangedSpecManifestPath logs\repair-committed-stimuli.json -OutputPath logs\repair-committed-plan.jsonnpm run test:ps -- -TestPath scripts\tests\evals\New-AgentEvalPlan.Tests.ps1vally linton the skill-behavior specAll 311 previous skill cases and grader identities remain intact. Six novel
advisory cases bring the suite to 317 cases across 85 subjects; each repaired
subject covers knowledge, generated output, and ownership/no-execution
boundaries. Two deterministic file-checker probe sets also passed. No model
trials or judge calibration were run as part of the local repair.
The full deep RPI repair loop completed through acceptance Review, with no
remaining required source finding. Hosted checks must rerun on the published
repair before CI recovery can be claimed.
Checklist
Required Checks
AI Artifact Contributions
hve-builderreview mode to review contributionhve-builderreviewRequired Local Checks
The following local-safe validation commands must pass before merging:
npm run validate:localnpm run validate:docsnpm run spell-checknpm run lint:md-linksSecurity Considerations
No dependencies, executable skill scripts, model settings, permission declarations, or agent tool configuration were changed.
GHCP Membership Changes
.github/instructions/shared/plain-language.instructions.md.Additional Notes
Human-only Builder, contribution-verification, and security attestations remain unchecked. The six new writing cases are advisory until calibrated. Installing the plugin or discovering an instruction is not proof that its contents reached a particular model request.