blast-radius-check
Skill8 starsChanged 6 months agoMITView on GitHub
Blast radius analysis - search for ALL usages before modifying any existing code. Prevents breaking unrelated consumers by scoping impact before scoping fix.
In this repository129 other files it gives its agents.
- AGENTS.mdAGENTS.md
- guidance/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- guidance/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- brainstormingSkill
- branch-flow-gateSkill
- branch-sync-gateSkill
- codebase-reconSkill
- codeowners-drift-auditSkill
- code-review-batterySkill
- cognitive-complexity-refactoringSkill
- debateSkill
- debug-conductorSkill
- dispatching-parallel-agentsSkill
- domain-buildSkill
- evidence-adjudicatorSkill
- executing-plansSkill
- external-cli-auditSkill
- feature-developmentSkill
- field-rename-verificationSkill
- finishing-a-development-branchSkill
- git-branch-conventionsSkill
- gitlab-cliSkill
- hotfix-charterSkill
- implementation-trackerSkill
- incident-revert-reviewSkill
- infra-config-investigatorSkill
- investigation-stateSkill
- kernel-splitSkill
- llm-behavior-investigatorSkill
- llm-skill-reviewSkill
- merge-authorization-gateSkill
- micro-harsh-reviewSkill
- output-verificationSkill
- pre-commit-gateSkill
- progressive-code-review-gateSkill
- progressive-harsh-reviewSkill
- providing-code-reviewSkill
- pr-triage-gateSkill
- push-authorization-gateSkill
- receiving-code-reviewSkill
- reproduction-experiment-investigatorSkill
- requesting-code-reviewSkill
- requirements-validation-pmSkill
- requirements-validationSkill
- scope-tripwireSkill
- session-handoffSkill
- skills-hierarchy-tuningSkill
- sp-bughuntSkill
- state-consistency-investigatorSkill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- test-driven-developmentSkill
- timeline-trace-investigatorSkill
- token-estimationSkill
- unified-commit-gateSkill
- using-git-worktreesSkill
- using-superpowersSkill
- verification-before-completionSkill
- The first 60 shown.
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 12 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈12
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 33
Before modifying any existing function, method, type, or API contract
- Distinctive: line 42
Before modifying ANY existing code, determine the full blast radius — every caller, consumer, and dependent that could be affected.
- Distinctive: line 73
If shared: Do ALL consumers need this change, or just ONE?
- Distinctive: line 74
If one: Should I modify the shared code, or create a new variant?
- Distinctive: line 116
Fill this table BEFORE implementing: Every empty row is a potential bug.
- Very specific: line 122
No TypeScript errors (field is optional or any)
- Distinctive: line 123
No test failures (tests mock inputs)
- Distinctive: line 124
No lint errors (code is valid)
- Distinctive: line 128
If yes → that transform must include new fields.
- Distinctive: line 132
After implementing, verify with cross-repo grep:
- Distinctive: line 168
Use the inventory from Step 1 as a checklist.
- Distinctive: line 181
After all edits, re-run the same grep commands from Step 1.
- Distinctive: line 194
CHECK 4c will fail if any structured file still references the deleted skill.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Blast Radius Check
Wrong skill? Pre-commit checks →
pre-commit-gate. Field renames →field-rename-verification. Output inspection →output-verification.Source:
superpowers-plusPart of: Engineering Rigor skill family
When to Use
- Before modifying any existing function, method, type, or API contract
- Refactoring shared utilities, base classes, or cross-cutting concerns
- Applying a "quick fix" or hotfix to production code
- Changing configuration, feature flags, or environment variables used by multiple services
The Rule
SCOPE YOUR IMPACT BEFORE YOU SCOPE YOUR FIX.
Before modifying ANY existing code, determine the full blast radius — every caller, consumer, and dependent that could be affected.
Why This Gate Exists
Common failure: Making changes without checking blast radius. The "fix" breaks multiple unrelated consumers because you only looked at the immediate problem, not the full dependency graph. Never assume a change is isolated without proving it.
Step 1: Search for ALL Usages
# Search ALL repos for the function/class/field you're modifying
grep -rn "functionName\|ClassName\|fieldName" --include="*.ts" .
# Check for imports
grep -rn "import.*functionName\|from.*moduleName" --include="*.ts" .
# Check for interface implementations
grep -rn "implements InterfaceName\|extends BaseClass" --include="*.ts" .Step 2: Categorize Dependents
| Dependent Type | Count | Impact if Changed |
| Direct callers | ||
| Subclasses/implementations | ||
| Test files | ||
| Config/types |
Step 3: Ask the Hard Questions
- Is this a shared utility (many consumers) or single-use code (one caller)?
- If shared: Do ALL consumers need this change, or just ONE?
- If one: Should I modify the shared code, or create a new variant?
Blast Radius Check
Wrong skill? Pre-commit checks →
pre-commit-gate. Field renames →field-rename-verification. Output inspection →output-verification.list
- How many files/functions call this code? (
grep -cto count) - Is this a shared utility or single-use code?
- Will this change require updates to OTHER components I'm not planning to touch?
- What's the worst-case impact if I get this wrong?
- Have I searched ALL repos, not just the one I'm working in?
Decision Matrix
| Blast Radius | Action |
| 1 caller | Safe to modify directly |
| 2-5 callers | Review each caller's usage before modifying |
| 5+ callers | Consider adding new function/parameter instead of modifying |
| Shared utility | Almost NEVER modify — extend or create variant instead |
If you can't answer "How many things use this?" — STOP and find out.
Data Flow Tracing
For multi-component changes, trace the full data flow BEFORE implementing:
Data/Control Flow Diagram:
SOURCE → STORAGE → ROUTER → CONSUMER → EXTERNALEnumerate ALL Touchpoints
| Layer | File(s) | Change Required |
| Schema | ||
| Storage | ||
| Router | ||
| Consumer | ||
| External |
Fill this table BEFORE implementing. Every empty row is a potential bug.
Identify "Silent" Pass-Through Points
These are files that receive data and pass it along WITHOUT using it locally. They're easy to miss because:
- No TypeScript errors (field is optional or
any) - No test failures (tests mock inputs)
- No lint errors (code is valid)
Question to ask: "Does this component TRANSFORM the data before passing it?"
If yes → that transform must include new fields.
Post-Implementation Verification
After implementing, verify with cross-repo grep:
# Search ALL repos for the new field/function
grep -rn "newFieldName" --include="*.ts" --include="*.js" repo1/ repo2/ repo3/
# Include snake_case, camelCase, UPPER_CASE variants
grep -rn "new_field_name\|NewFieldName\|NEW_FIELD_NAME" .Skill Deletion Protocol
Deleting or renaming a skill is a high blast-radius operation. Skills are referenced in many forms across many file types. Run this exact sequence before any deletion or rename.
Step 1: Full Reference Inventory (run first — before any edits)
SKILL="skill-name" # e.g., engineering-rigor
MERMAID="${SKILL//-/_}" # e.g., engineering_rigor
# All text files — hyphenated skill name
grep -rn "$SKILL" . --include="*.md" --include="*.sh" \
--include="*.js" --include="*.json" --include="*.yaml" 2>/dev/null
# Mermaid node and edge declarations — underscore form
grep -rn "$MERMAID" . --include="*.md" 2>/dev/null
# Structured data files — these FAIL harsh-review if stale
grep -n "\"$SKILL\"" tools/high-cost-skills.json \
docs/composition-manifest.json 2>/dev/nullBuild a complete list of every hit before touching a single file. If the count is higher than expected, that's the real blast radius.
Step 2: Edit All Files
Use the inventory from Step 1 as a checklist. Touch every file on the list — including:
| Location | What to check |
skills/*/skill.md | coordination.requires, .enables, .escalates_to |
tools/high-cost-skills.json | Array entries |
docs/composition-manifest.json | Object keys |
docs/skill-dependency-graph.md | Node declarations AND edge declarations (underscore form) |
lib/intent-patterns.js | Skill name strings |
docs/*.md prose | Companion skill tables, "Wrong skill?" headers |
Step 3: Re-run the Identical Grep
After all edits, re-run the same grep commands from Step 1. Zero hits = done.
grep -rn "$SKILL\|$MERMAID" . --include="*.md" --include="*.sh" \
--include="*.js" --include="*.json" 2>/dev/null | grep -v CHANGELOGStep 4: harsh-review.sh
bash tools/harsh-review.shCHECK 4c will fail if any structured file still references the deleted skill. A passing gate is mandatory before commit.
Companion Skills
pre-commit-gate— Before committing changesproviding-code-review— When reviewing others' PRs- autonomous-chain-controller: Chain-aware refactoring
Failure Modes
| Failure | Fix |
| Only checked direct callers, missed transitive consumers | Trace data flow through ALL paths (READ → STORE → PASS) |
| Skipped test impact analysis | Run full test suite, check for tests that exercise changed paths |
| Assumed internal function has no external consumers | Grep for ALL references — internal/external distinction is often wrong |
| Changed API contract without checking client services | Use field-rename-verification for cross-service contract changes |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

