You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Add complete product-design-harness skill for validated coding handoffs - #5321
I have read the CONTRIBUTING guidelines; the missing Copilot inference test is explicitly disclosed below.
I have read the guidance for submissions involving paid services. This skill is an MIT local bundle, with no required paid service or signup.
My contribution adds a new skill in the correct directory.
The folder and skill name follow the required naming convention.
The content is structured and includes bundled schemas, working examples, and validation instructions.
I have tested the skill with GitHub Copilot.
I have run npm run build (the target of npm start) and verified generated documentation is up to date.
I am targeting the main branch.
Description
Adds product-design-harness for reviewing a product change before a coding agent implements it. Its concrete handoff constraint is that the context pack's scope must exactly equal the reviewed execution_boundary.may_do; a continue verdict does not authorize expanding that boundary. The bundle supplies the JSON schemas, examples, and a deterministic joint review/context-pack validator rather than relying on a generic instruction to review carefully.
For example, a Copilot user can review an onboarding change, validate the supplied review and context pack together, and stop implementation if an extra action appears in the handoff scope. The workflow and command are in resources/docs/LIVE-CODING.md. The source also supports verify and stop_reframe outcomes; this submission does not add agent permissions, hooks, MCP servers, or a hosted service.
Type of Contribution
New skill file with its complete supporting folder.
All 91 files are included, including the original LICENSE, Python validators, schemas, examples, prompts, knowledge, documentation, and both SVG diagrams. Relative to that source, SKILL.md only quotes the existing YAML description with single quotes. Two bundled schema files replace [Pp]roduction with the equivalent (?:Production|production) so codespell does not misread the regex suffix as a typo. All other 88 supporting files, including MIT LICENSE, are byte-identical, and the original Markdown body is unchanged. The only generated catalog change is one row in docs/README.skills.md.
Validation and limitations
Validation binds candidate commit c9e35bbd8e016b0c9546f64a4a0d9e5b89126e3e, based on d85f762d6e521abf5fe16703946e9b38adc42da7:
Official npm run skill:validate: baseline 442 skills passed; candidate 443 skills passed.
Official npm run build: baseline and candidate passed; only the expected skill catalog row changed. npm run plugin:validate also passed. Dependencies were installed from the unchanged lockfile with lifecycle scripts disabled.
For each of the two changed schema patterns, 171 cases covered production/Production/PRODUCTION, all other alternatives, substring matching, mixed case, and non-matches. Python and ECMAScript regex results agreed between the original and adapter; both complete JSON schemas also produced identical validation errors across the same 171 cases each. The replacement changes no other pattern text or matching scope.
Existing gh 2.93.0 native gh skill install installed this fork's exact candidate SHA, selecting skills/product-design-harness, into an isolated local directory. All 91 files were present with no missing/extra files; all 90 supporting files and MIT LICENSE were byte-identical to the submitted bundle.
gh rewrote SKILL frontmatter and injected its repository/path/ref/pin/tree provenance. Original frontmatter values were preserved; it also normalized one leading blank line after frontmatter. These installer changes are not claimed as byte-identical SKILL output.
From the installed resources directory, the existing joint validator accepted examples/live-coding-review.json with examples/live-coding-context-pack.json (exit 0). It rejected an isolated context-pack copy with an extra unapproved scope action (exit 1). Validation used existing Python/jsonschema 4.26.0.
No GitHub Copilot model session or other inference was run, so that checklist item remains unchecked. Native file installation and deterministic contract validation do not establish model effectiveness, security certification, maintainer acceptance, or an existing public catalog listing. Any later host/provider usage follows the user's own environment and authorization. No additional CLA or separate rights grant has been signed as part of this submission.
Observed PR CI limitation
The original candidate f4d9ff922d29b57e27d84b8c2b607f80a93b0d04 passed official Skills and README validations, line-ending and license checks, but codespell failed on roduction in two unchanged upstream regexes. The current candidate makes only the two equivalent schema spelling adaptations described above, with acceptance/rejection evidence; no ignore list, skip configuration, workflow, or gate was changed. Current exact-head CI and maintainer acceptance must be evaluated separately; this PR does not claim all CI is green.
By submitting this pull request, I confirm that my contribution abides by the Code of Conduct and will be licensed under the MIT License.
Risk tier:merge-risk:high — Privileged execution, automation, or review-policy change Required to merge: passing submission-gate checks plus 1 approval from reviewers with write access, including a maintainer with admin or maintain permission.
Why this tier
skills/product-design-harness/resources/scripts/check_override.py is a high-risk path (automation, scripts, MCP config, hooks, or review policy)
skills/product-design-harness/resources/scripts/check_review.py is a high-risk path (automation, scripts, MCP config, hooks, or review policy)
skills/product-design-harness/resources/scripts/jev_gate.py is a high-risk path (automation, scripts, MCP config, hooks, or review policy)
Label needs-review:HIGH flags a high contributor-risk signal
| 3 | Install with `npx skills add cis2042/ux3-product-design-harness -g -y`, or copy `skills/product-design-harness/SKILL.md` together with its `resources/` directory into the age
| 3 | Install with `npx skills add cis2042/ux3-product-design-harness -g -y`, or copy `skills/product-design-harness/SKILL.md` together with its `resources/` directory into the age
✓ [spec-compliance] All 1 skill(s) are spec-compliant.
ℹ️
✓ spec-compliance: All spec checks passed.
ℹ️
✓ [valid-refs] All file references across 1 skill(s) are valid.
ℹ️
✓ valid-refs: All file references resolve to existing files within the skill directory.
ℹ️
1 skill(s) linted, 1 passed
Full linter output
### Linting skills/product-design-harness
✅ product-design-harness (2/2 checks passed)
✓ [spec-compliance] All 1 skill(s) are spec-compliant.
✓ spec-compliance: All spec checks passed.
✓ [valid-refs] All file references across 1 skill(s) are valid.
✓ valid-refs: All file references resolve to existing files within the skill directory.
1 skill(s) linted, 1 passed
Please assign a maintainer for source review of this submission. This is a review request, not a claim that the skill is accepted or ready for release.
Known upstream issues remain in the current candidate: surrounding whitespace can bypass the override forbidden-action comparison, and three SVG references are missing from the installed skill directory. Fixes are proposed in cis2042/ux3-product-design-harness#23 and #24, but are not merged. I will not represent those fixes as present in this candidate; the source needs updating and revalidation before acceptance. Copilot model/task testing has not been performed.
Thanks @cis2042 for the very transparent write-up. We're going to pass on this one.
The genuinely useful part for Copilot users is narrow: the rule that a coding handoff's scope must exactly equal the reviewed execution_boundary.may_do, enforced by check_review.py. But it arrives inside a 92-file product-strategy framework whose own README describes PMs and designers as the audience. Most of SKILL.md is practice a capable model already follows, and pre-implementation product framing is already covered by prd and breakdown-feature-prd. We'd also need the check_override.py bypass and missing SVGs fixed, real Copilot testing, and clarity on the rights to the UX3 curriculum material.
If you'd like to contribute the handoff-scope check on its own — a short skill plus the validator, tested in Copilot — we'd be happy to look at a new PR.
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
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.
github-actions
Bot
added
new-submission
Pull Request Checklist
npm run build(the target ofnpm start) and verified generated documentation is up to date.mainbranch.Description
Adds
product-design-harnessfor reviewing a product change before a coding agent implements it. Its concrete handoff constraint is that the context pack's scope must exactly equal the reviewedexecution_boundary.may_do; a continue verdict does not authorize expanding that boundary. The bundle supplies the JSON schemas, examples, and a deterministic joint review/context-pack validator rather than relying on a generic instruction to review carefully.For example, a Copilot user can review an onboarding change, validate the supplied review and context pack together, and stop implementation if an extra action appears in the handoff scope. The workflow and command are in
resources/docs/LIVE-CODING.md. The source also supports verify and stop_reframe outcomes; this submission does not add agent permissions, hooks, MCP servers, or a hosted service.Type of Contribution
Source, completeness, and license
I am the upstream publisher. The fixed MIT source is cis2042/ux3-product-design-harness at 7f9f724c5e31813d1c35767f9f454fe9a1e96891.
All 91 files are included, including the original LICENSE, Python validators, schemas, examples, prompts, knowledge, documentation, and both SVG diagrams. Relative to that source, SKILL.md only quotes the existing YAML description with single quotes. Two bundled schema files replace
[Pp]roductionwith the equivalent(?:Production|production)so codespell does not misread the regex suffix as a typo. All other 88 supporting files, including MIT LICENSE, are byte-identical, and the original Markdown body is unchanged. The only generated catalog change is one row indocs/README.skills.md.Validation and limitations
Validation binds candidate commit
c9e35bbd8e016b0c9546f64a4a0d9e5b89126e3e, based ond85f762d6e521abf5fe16703946e9b38adc42da7:npm run skill:validate: baseline 442 skills passed; candidate 443 skills passed.npm run build: baseline and candidate passed; only the expected skill catalog row changed.npm run plugin:validatealso passed. Dependencies were installed from the unchanged lockfile with lifecycle scripts disabled.gh skill installinstalled this fork's exact candidate SHA, selectingskills/product-design-harness, into an isolated local directory. All 91 files were present with no missing/extra files; all 90 supporting files and MIT LICENSE were byte-identical to the submitted bundle.resourcesdirectory, the existing joint validator acceptedexamples/live-coding-review.jsonwithexamples/live-coding-context-pack.json(exit 0). It rejected an isolated context-pack copy with an extra unapproved scope action (exit 1). Validation used existing Python/jsonschema 4.26.0.No GitHub Copilot model session or other inference was run, so that checklist item remains unchecked. Native file installation and deterministic contract validation do not establish model effectiveness, security certification, maintainer acceptance, or an existing public catalog listing. Any later host/provider usage follows the user's own environment and authorization. No additional CLA or separate rights grant has been signed as part of this submission.
Observed PR CI limitation
The original candidate
f4d9ff922d29b57e27d84b8c2b607f80a93b0d04passed official Skills and README validations, line-ending and license checks, but codespell failed onroductionin two unchanged upstream regexes. The current candidate makes only the two equivalent schema spelling adaptations described above, with acceptance/rejection evidence; no ignore list, skip configuration, workflow, or gate was changed. Current exact-head CI and maintainer acceptance must be evaluated separately; this PR does not claim all CI is green.By submitting this pull request, I confirm that my contribution abides by the Code of Conduct and will be licensed under the MIT License.