Prism before the edit · Shale after it

Every AI-written change should explain itself.

Prism shows what a change can affect before code is touched. Shale records what was intended, changed, and checked so the pull request carries its own evidence. Mason is the reference agent that uses both.

Explore Prism Explore Shale

Two tools for the same code change.

Prism helps the work begin with complete context. Shale helps review begin with an honest record.

before the edit · core project

Prism

Ask a change-shaped question and get the declaration, implementations, callers, tests, and concrete edit sites in one deterministic result.

How Prism works

after the edit · core project

Shale

Carry intent, touched files, observed checks, integrity signals, and explicit gaps with the code into pull-request review.

How Shale works

Mason

The incubating reference agent. It puts Prism change intelligence and Shale evidence into the harness so the complete loop can be measured end to end.

View Mason

Know the whole change before you touch it.

Prism turns the multi-step investigations developers and agents repeat every day into task-shaped operations with explicit completeness.

prism · change impact

$ prism change-impact 'ResponseWriter.Status'

declaration     server/response.go:18
family          3 implementations
callers         8 resolved call sites
declaringTypes  1 interface block

completeness: closed

change-impact

Everything affected by a method signature change, including resolved callers text search misses.

rename-plan

The same complete set converted into concrete, drift-checked line edits.

verify

Every site an agent-authored diff missed — line-precise, exit 1 if incomplete.

Get Prism See the benchmarks

Make the result explain itself in review.

Shale turns the agent session into a compact evidence card that travels with the code. It shows what is known and what is missing.

A Shale session: intent, edits, checks, completion, and evidence publication

github-actions bot commented

Shale · 1 session · claude-code

60k tokens · 2 iterations · < 1 min

Intent

Add rate limiting to the login endpoint
Token bucket per client IP. Return 429 with Retry-After.

Changed files (3) · all with evidence

`ratelimit.go`	verified
`ratelimit_test.go`	verified
`main.go`	verified

Checks recorded locally

go test ./... passed

Advisory · CI is authoritative.

The evidence format is real. See the live demo pull requests.

Get Shale Read the Shale guide

Local tools. Open evidence.

Each project works on its own and fits the tools developers already use. There is no shared account or hosted control plane.

Local first

Code analysis and session capture stay with the repository and developer workflow.

Measured claims

Correctness results are tied to published tasks, independent oracles, and raw runs.

Honest boundaries

Unsupported analysis and missing evidence are labeled instead of silently promoted to certainty.

Under Prism: Grove. Prism embeds Grove's semantic code graph. Grove remains available to tool builders, but Prism users do not install or configure it separately.