Skip to content
Agentmart
Submit a tool

blast-radius

Trace what a change could break beyond its diff and prove the load-bearing safety assumptions through real code.

Skill by matthewyjiang

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/matthewyjiang-blast-radius

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Trace what a change could break beyond its diff and prove the load-bearing safety assumptions through real code. Use before risky changes or shipping.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/blast-radius

What's inside

Skill name
blast-radius
Description
Trace what a change could break beyond its diff and prove the load-bearing safety assumptions through real code. Use before risky changes or shipping.
License
MIT
Instructions
30 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Blast radius

Read [the workflow contract](../../references/workflow-contract.md).
`how` explains mechanism; `why` explains its shape. This skill finds what breaks
elsewhere, including dependencies a symbol search cannot see.

1. Pin the actual diff and intent; include deleted/renamed contracts and implicit
   behavior. Pull relevant historical/PR evidence when it explains a safety invariant.
2. Find the load-bearing safety facts. A long caller list is not a risk analysis.
3. Trace beyond grep: pinned library implementations/local patches, wire formats,
   storage schemas, feature gates, lifecycle ordering, teardown, tasks and other
   languages/services consuming the same data. Never invent a caller or API.
4. Distinguish confirmed risks, cleared cases and unknowns. Give a reachable failure
   scenario, likelihood/impact rationale and file/line evidence, not guessed percentages.
5. Prove each important fact through the real code or running user path. A small
   script importing the shipped library often beats a persuasive explanation.
   Parent runs validation; no delegate builds. If execution needs unavailable
   credentials/hardware or would have unsafe effects, mark it **unproven**.
6. For broad/high-risk review, an independent `reviewer` pass can pressure-test the
   assumptions. Use `interrogate` for cross-model input when model selection is
   available. Do not block ordinary proof on model diversity or duplicate PR review:
   `thermos` remains the single combined PR review entry point.

Report: what changed; the facts safety depends on; proof artifacts/commands and
observed results; risks with evidence; cleared cases; cheapest meaningful check
before merge. State evidence strength explicitly: claim, source trace, demonstrated
unreachable path, real-code execution, or running-app reproduction. Do not present
source reading as runtime proof. Keep private material out of public artifacts.

From matthewyjiang/rho-pstack, MIT.

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Oct 17, 2026 (12 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from matthewyjiang

Cite this page

blast-radius skill. Agent Mart. https://agentmart.dev/skills/matthewyjiang-blast-radius (last updated 2026-10-24)