Miller impact analysis
Use when assessing blast radius with Miller, choosing tests for a change, answering who uses a symbol, or planning a refactor.From its SKILL.md
npx -y skills add anortham/miller --skill miller-impact-analysisAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.0 KB, as published. Nobody here has run it
Miller Impact Analysis
Use Miller's graph-backed impact tool before refactors or shared behavior changes. One impact call should replace manual reference greps.
Workflow
- Resolve the target if it is ambiguous:
search(query="<symbol or concept>")
inspect(target="<candidate>")
- Run exactly one impact seed:
impact()
impact(target="<symbol-or-file>")
impact(changed_paths=["<file1>", "<file2>"])
impact(diff="<unified diff>")
With no args, impact() reads the working-tree git diff and maps changed ranges to impacted symbols plus likely
tests — run it after edits, before committing, to see what your uncommitted change affects. Use git=true
(base/staged imply git) to scope to a specific git diff instead.
- Inspect or trace only the risky results:
inspect(target="<impacted-symbol>", depth="full")
trace(target="<impacted-symbol>")
trace(target="<source>", mode="path", to="<sink>")
- If
impactreports likely tests, treat that list as the starting verification set. Add broader tests when the touched surface is shared infrastructure or cross-workspace behavior.
Risk Tiers
- High: public entry points, tool handlers, workspace/indexing paths, shared contracts, or many downstream symbols.
- Medium: feature modules, storage/read-path code, bridge providers, or several callers.
- Low: isolated helpers, tests, docs, or one-off internal callers.
Report
Lead with:
- Target and definition
- High/medium/low risk groups
- Likely tests or verification commands
- Unknowns such as stale index, ambiguous symbol names, or dynamic usage
Do not replace impact with rg unless Miller cannot resolve the target.