Panda Suite Docs
v1.9.2GitHub ↗
Panda Suite / Panda Impact

Panda Impact

Examples use standalone Codex names. For a Codex plugin, use $panda-suite:panda-impact; for a Claude Code plugin, use /panda-suite:panda-impact. See all invocation formats.

Panda Impact assesses a proposed change before it is implemented: which components depend on it, what business behavior and stored data it touches, the rollout risks, and the verification it needs. It is read-only by default.

At a glance

Topic Summary
Use it for API or schema changes, renaming fields, replacing a library, moving a module, changing an event shape
Not for A defect investigation (Debugger), a migration execution plan (Plan), an ordinary diff review (Review)
Edits by default None, apart from an analysis document you request
Main output Confirmed dependencies, affected components, data/API/behavior impact, operational risks, compatibility assessment, mitigations, verification, unknowns

Confidence labels

Label Meaning
Confirmed impact A demonstrated dependency plus an incompatible change
Likely impact Evidence fits, but a key boundary is unverified
Possible risk A plausible condition that has not been tested
Unknown Evidence cannot decide

Each finding states its consequence, trigger, evidence, missing check, and priority. A matching identifier alone does not prove an affected runtime path.

How it works

  1. Bound the change. Old and new contract, affected versions and environments, expected benefit, scope. The target stays "proposed" until approved.
  2. Trace dependencies. Producers and consumers through callers, models, queries, serializers, UI, jobs, events, integrations, configuration, and tests, with precise paths. Each negative search is bounded by its scope.
  3. Assess boundaries. Stored-data compatibility, request/response and event contracts, validation and errors, deployment ordering, mixed versions, background work, monitoring, and recovery. Code-defined schema is kept separate from observed database contents.
  4. Classify findings with the labels above.
  5. Recommend mitigation. The least complicated compatible approach, with tests, consumer coordination, migration and rollout checks, observability, and rollback limits. A lossy conversion may need a backup restore rather than a reverse script; reversibility is never promised without evidence.
  6. Validate coverage. Every material impact has a mitigation or explicit decision, and uninspected repositories, consumers, or data are named.

What to provide

  • The exact proposed change (old and new shape).
  • The repositories, contracts, or schemas available for tracing.
  • Known external consumers, if any.

Examples

Change a field's type

$panda-impact Analyze replacing order_status strings with integer codes. Trace
models, queries, API responses, UI callers, jobs and tests that actually exist.
Separate confirmed dependencies from possible external consumers and unknowns.
Assess compatibility, migration/recovery risks and required verification. Read-only.

Rename a public API field

$panda-impact We want to rename "customerId" to "accountId" in the public
REST API. What breaks, which clients might be affected, and what compatibility
options exist (aliasing, versioning)? Read-only.

Split a table

$panda-impact Assess splitting the users table into users and profiles.
Include ORM models, raw SQL, reports, background jobs and fixtures. Note
deployment ordering for mixed application versions.

Replace a dependency

$panda-impact What is affected if we replace moment.js with date-fns? List every
call site and any behavior differences (time zones, parsing) that could change output.

Change an event payload

$panda-impact We plan to drop the "legacy_total" field from the OrderPlaced
event. Trace producers and consumers in this repository and list the consumers
we cannot see.

Expect "Unknown" for consumers outside the inspected repositories, with the smallest observation that would resolve it.

Configuration default change

$panda-impact What happens if we change the default session timeout from 30
days to 7 days? Include mobile clients, remember-me tokens and support load.

Save the assessment

$panda-impact Assess moving file uploads from local disk to object storage.
Save the analysis to docs/upload-storage-impact.md with a small dependency diagram.

Boundaries

Impact does not edit source, run migrations, modify data, or trigger deployment or external effects unless separately authorized. Missing consumer or production access lowers confidence; it never proves that no consumer or stored record exists.

  • Plan to turn the mitigations into tasks and release gates
  • Spec to define the new contract precisely

Skill instructions

Read the full Panda Impact instructions in the repository.