Help us improve
Share bugs, ideas, or general feedback.
From tap-skills
Analyze the impact surface of a PR or set of changes before merging. Use when someone says "blast radius", "review this PR", "what does this change affect", "is this safe to merge", "impact analysis", or after an agent opens a PR that needs human verification. Maps what changed, what else is affected, what could break, assigns risk level, and generates a manual verification checklist. The human gate for mixed human-agent teams.
npx claudepluginhub teambrilliant/marketplace --plugin tap-skillsHow this skill is triggered — by the user, by Claude, or both
Slash command
/tap-skills:blast-radiusThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Map the impact surface of changes. Help humans focus their limited attention on what matters before merging.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Share bugs, ideas, or general feedback.
Map the impact surface of changes. Help humans focus their limited attention on what matters before merging.
Determine what to analyze (in priority order):
gh pr diff <number>git diff main...<branch>git diff main...HEAD (current branch vs main)Also read:
.tap/tap-audit.md for environment URLs (needed for verification checklist)In 1-2 sentences, state what the change is TRYING to do. This frames the entire analysis — every finding relates back to "does this serve or endanger the stated intent?"
For each changed file, trace the impact outward:
Direct changes:
Dependents (ripple effects):
grep / codebase search to find callers and importersShared state:
Test coverage gaps:
LOW — Merge confidently
MEDIUM — Test these specific flows
HIGH — Test everything on the checklist
Surface what the static analysis CAN'T see. Scan the diff for patterns that indicate hidden dependencies, grounded in Ousterhout's complexity symptoms:
Obscurity — behavior depends on information not visible in the diff:
Hidden dependencies — connections not traceable via static imports:
Change amplification risk — external consumers of changed interfaces where the full set of consumers is unknown:
Each blind spot states: [what's hidden] — [why it matters for this PR].
Merge blind spots that overlap with Verification Checklist items into the checklist instead of duplicating. Remaining blind spots feed additional checklist items.
Skip this section entirely for LOW risk PRs.
Produce a concrete list of things to manually test. Be specific — pages, flows, inputs, edge cases.
Format:
□ [page/flow] — [what to verify] — [why it might break]
Include:
.tap/tap-audit.md has URLs, reference specific environmentsPrioritize the checklist: most likely to break first, most damaging if broken second.
Call out anything that doesn't look right:
Use the template in references/blast-radius-template.md.
Human mode: Interactive walkthrough. Start with risk level and intent summary. Walk through impact map. Present verification checklist. Ask: "Want me to dig deeper into any area?"
Agent mode: Write structured report as a PR comment via gh pr comment. Include risk level, impact map, and verification checklist. Human reads async.