Our take Editorial verdict
Blast Radius answers the question agents usually hand-wave: what else could this change break? It traces dependents and callers, then requires the one safety-critical assumption to be proven by running code rather than asserted in prose.
Best for
Small-looking changes to shared code, public interfaces, data shapes, or configuration where the real risk lives downstream.
Keep in mind Limits and tradeoffs
Tracing has diminishing returns in weakly typed or dynamically wired codebases, and the extra verification step is exactly the kind of thing a rushed agent will skip unless you ask for it explicitly.
How to trace the downstream risk of a change
- Provide the proposed interface, field or configuration change and the relevant repository context.
- Ask the agent to trace callers, readers and writers and identify the safety-critical assumption.
- Require a runnable check of that assumption and inspect the affected boundaries before accepting the change.
This is a source-based editorial review. No retained execution or installation test is published here. The starter request is illustrative, not an observed result. Check the setup and compatibility caveats before using it.
Explore Lauren Tan’s Pstack · Skill reference and original instructions
Try this skill
Install · Collection / plugin
Run in Claude Code chat. Installs the whole Pstack plugin, including its other playbooks, principles and agents. Multi-model workflows may require paid model access. Claude Code skips Cursor-only features. Scope is chosen by the plugin manager.
/plugin marketplace add cursor/plugins /plugin install pstack@cursor-plugins
Command / route checked 30 October 2026. Upstream installation
Setup requirements and caveats
Run in Claude Code chat. Installs the whole Pstack plugin, including its other playbooks, principles and agents. Multi-model workflows may require paid model access. Claude Code skips Cursor-only features. Scope is chosen by the plugin manager.
These are documented requirements; installation has not been tested here.
Your agent needs its own account, runtime and model access. Usage charges depend on your plan; unrecorded costs are unknown.
Run the example
Start a fresh session in the target project. Provide the relevant files or diff. This is the review’s example request, not a record of your execution.
Use the blast-radius skill. Before we rename this field, map every reader and writer and prove with a test run that nothing outside the module depends on the old shape.
Check your result · Successful first run
A map of field readers and writers plus evidence of the rename’s affected boundaries.
Confirm the agent loaded the named skill, inspect the output and run the relevant checks. Installation alone does not prove a successful run.
Something not working?
The skill is missing from my agent
Check that the skill is installed for the selected agent and project/global scope. Open a fresh session in that project and check the agent’s skill list. Plugin installs may also need to be enabled.
The wrong skill runs, or nothing happens
Start a fresh conversation, explicitly name the skill using the request above, and provide the requested inputs. Ask the agent to confirm which skill it loaded. Check for duplicate installed copies or overlapping instructions.
The skill runs but the result is poor
Compare the result with the successful-first-run check above. Supply the missing files, reproduction steps or constraints, then ask for one focused correction. If the skill addresses a different problem, return to the task guide and compare alternatives.
Installation troubleshooting · missing dependencies
If npx fails, check the Node/npm versions in setup requirements, Git and network access. For skills CLI installs, run npx skills list (add -g for global). If a route changed, follow upstream instructions.
Collection background and tradeoffs
It is Cursor-first and uses plugin and multi-model capabilities. Claude Code has an official adapter that skips Cursor-only features; other agents require separately evaluated community ports. The strong opinions, large playbook surface, and autonomous-run patterns need judgment to apply well, and the collection deliberately omits planning skills because it prefers a specification expressed as code.
Source maintained by Lauren Tan (poteto). Read original skill. Compatibility guidance and published showcase runs remain separate from your first run.
Evidence and provenance Sources and retained files
Evidence and provenance
Review method not recorded. The editorial date does not establish a source inspection or test run.
No executed test recorded. Compatibility guidance and source review do not establish tested behavior.