♻️ Make the Vizzly skill portable#318
Merged
Merged
Conversation
Start agents with the bounded evidence handoff, then point them to real images, returned commands, and opt-in Honeydiff detail. This keeps the packaged guidance aligned with the CLI contracts that now ship.
Center the packaged skill on the portable Agent Skills contract, factual CLI workflows, and capability-based evidence inspection. Remove unsupported product guidance and keep OpenAI metadata as an optional adapter.
Robdel12
changed the title
Keep the portable skill focused on the intended workflow. The CLI will own emitting directly executable structured context commands in a separate change.
Robdel12
marked this pull request as ready for review
July 22, 2026 04:48
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
The packaged Vizzly skill is meant to work across skills-compatible agent harnesses, but its default workflow was not portable or consistently correct. The primary local command omitted
--json, so agents missed the bounded evidence payload; the TDD example shell-backgrounded a daemon that already detaches; and the references included incomplete parallel CI, overstated screenshot identity, and unsupported Public Properties guidance.Approach
Center the skill on the open Agent Skills shape: one lean, harness-neutral evidence loop with detailed CLI, SDK, dynamic-content, and setup workflows loaded only when needed. Agents now use the repository's existing CLI invocation, inspect images through whatever capability their harness provides, keep inaccessible or missing evidence unknown, and require explicit authorization before setup or review mutations.
The OpenAI YAML remains optional interface metadata. It defines no behavior, and the reusable contract lives entirely in
SKILL.mdand its references. The unverified public-screenshot guide is removed instead of presenting dashboard behavior as a supported CLI workflow.Evidence
Two independent audits checked portability and every documented command against the CLI source. Fresh forward tests correctly kept a terminal-only local investigation offline and used one shared ID plus one finalizer for a four-shard CI build. The skill validator and the full release suite pass, and the packed npm artifact contains exactly the portable skill, four focused references, and the optional adapter.