show-me

Generates local HTML evidence decks and visual walkthroughs from verification, review, and demo artifacts.

10|Updated May 12, 2026
One-click install
npx skills add https://github.com/sandsower/beislid --skill show-me-sandsower
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: show-me
Source: https://github.com/sandsower/beislid/tree/main/skills/show-me
Command: npx skills add https://github.com/sandsower/beislid --skill show-me-sandsower

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Chat and terminal output are poor mediums for proving a change works, walking through code, or reviewing a subsystem visually. This Skill turns verification results, diffs, media, logs, and sourced facts into a polished local HTML portfolio or deck with clear evidence status and provenance. ## Core Features & Use Cases - Evidence decks with explicit status: Classify output as verification, review, code-walkthrough, UI/CLI demo, docs, understanding, or mixed, with statuses like PASS, FAIL, INCOMPLETE, NOT SHOWN, and NEEDS CAPTURE so missing evidence is never hidden. - Structured portable document model: Produces index.html, show-me.json, manifest.json, copied media assets, and command logs with provenance (branch, commit, timestamps, exit codes, redactions). - Optional Lavish visual routing: After the canonical deck is rendered, it can route to a Lavish visual surface per repo config (suggest/prompt/auto) with non-fatal fallbacks; the portable deck always remains canonical. - Use Case: After implementing a feature, ask for a proof deck: the Skill gathers test logs, screenshots, and diffs, then renders a dark-styled HTML report showing exactly what passed, what failed, and what still needs capture. ## Quick Start Ask the agent to create a show-me evidence deck proving the current branch's changes work, including test output and screenshots.

Frequently Asked Questions about show-me

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a visual evidence deck for a code change?▼

Ask the agent to generate a show-me deck for the change. It classifies the intent, gathers diffs, test logs, and media, then renders a local index.html with show-me.json, provenance, and explicit PASS/FAIL/INCOMPLETE status per claim.

What output files does the show-me deck produce?▼

It produces an index.html report, a portable show-me.json document model, a manifest.json, copied image/video/GIF/diagram assets, and command logs under a timestamped directory in the Beislið state dir or a repo-local .beislid/show-me folder.

Can show-me run automatically inside verification workflows?▼

No. In v1 it is manual and user-requested only, enforced by a hard gate. Other Beislið workflows may suggest a visual deck, but they must wait for the user to explicitly ask before generating one.

Does show-me require the Lavish plugin to work?▼

No. Lavish routing is optional and only activates when repo config defines beislid:visual_surfaces with a non-off mode. Missing binaries, declined prompts, or command failures fall back non-fatally to the normal portable HTML deck.

What happens when evidence like screenshots is unavailable?▼

The deck adds a visible NOT SHOWN or NEEDS CAPTURE block instead of implying the evidence exists. Failure decks with FAIL or INCOMPLETE status are valid output and include suggested next checks.

How does show-me handle secrets and sensitive media?▼

Text logs, JSON, and HTML receive best-effort secret redaction noted in provenance. Screenshots and videos carry media-sensitivity warnings unless a real redaction tool was used, and redaction is never claimed falsely.