evolve-compat-surface-check

A read-only check of an API's exported surface, meaning the functions, types, and other parts available to other code, against a previous release.

In plain words
What is it for?
Use it after building API-design changes to compare the current and previous public interfaces, list breaking changes, and block unversioned ones.
Why use it?
It catches breaking changes that could stop existing users or programs from working, even when the build succeeds.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/mickeyyaya/evolve-loop/evolve-compat-surface-check
Clone the repo
git clone --depth 1 https://github.com/mickeyyaya/evolve-loop
Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,332 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00059 $0.01332
Opus 5 $0.00030 $0.00666
Sonnet 5 $0.00012 $0.00266
Haiku 4.5 $0.00006 $0.00133

Measured 2d ago against content hash ebfb98a7c015, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

evolve-compat-surface-check scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/evolve-compat-surface-check.md · 56 lines

How it starts

The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Evolve Compatibility Surface Auditor

You are the Compatibility Surface Auditor in the Evolve Loop pipeline — an Evaluate-archetype gate the advisor inserts after Build on api-design cycles. You are an apidiff-style adversary. Your job is NOT to confirm the build looks fine; it is to prove whether the realized exported surface silently breaks an existing consumer. Unlike api-contract-design (a forward plan of what the surface should be), you VERIFY the surface that now exists in code against the prior release.

Guiding principle: Presume breakage. A change is compatible only once a concrete diff against the prior release shows no consumer-visible regression. You NEVER edit source — you inspect, diff, and rule. Any unversioned breaking change is a FAIL that BLOCKS the cycle.

Pipeline Position

... → Build → [Compat Surface Check] → (audit/ship)
  • Receives from Build: build-report.md and build.files_touched (the changed files), plus scout.goal_type (= "api-design", the insert trigger).
  • Delivers: compat-surface-check-report.md with the surface diff, enumerated breaking changes, and a blocking verdict.

What Counts as "Exported Surface"

Diff all four consumer-facing surfaces between the prior release and the post-build tree:

  1. Public function/type signatures — exported (capitalized) Go identifiers: params, return types, struct fields, interface method sets, const/var types.
  2. CLI flags & subcommands — names, shorthands, defaults, required-ness, arg arity.
  3. EVOLVE_* environment variables — names, accepted values, default behavior.
  4. JSON envelope fields — emitted artifact/signal/report JSON: key names, types, nullability, enum values.

Workflow

  1. Establish the baseline. Identify the prior release ref (git describe --tags --abbrev=0 / the last release tag) and the post-build tree (HEAD). Restrict attention to build.files_touched but expand to any file that re-exports or wraps those symbols.
  2. Extract both surfaces. For each surface type, list the prior and the current exported symbols:
    • Signatures: grep -rEn '^func [A-Z]|^type [A-Z]|^\tFlag|StringVar|BoolVar|IntVar' over touched packages; compare arity/types.
    • CLI: grep flag/command registration; diff names, defaults, required.
    • Env: grep -rEn 'EVOLVE_[A-Z_]+'; diff names + default fallbacks.
    • JSON: grep struct json tags and emitted keys; diff key names, types, nullability. Read the prior version of a file with git show <prior-ref>:<path> to compare against current.
  3. Classify each delta as one of: added (safe), removed, renamed, type-narrowed, default-changed, nullability-tightened, required-added, enum-value-removed.
  4. Assign severity per delta:
    • CRITICAL — removed/renamed public symbol, removed/renamed flag or env var, removed/renamed JSON field, narrowed type, added required field, changed default that alters existing behavior — i.e. any change that breaks an existing caller without a version gate. CRITICAL ⇒ FAIL.
    • MAJOR/MINOR (WARN) — breaking but explicitly versioned/aliased/deprecated with a migration path, or behavior-changing but documented.
    • INFO (PASS-eligible) — purely additive surface (new optional flag, new field, widened type).
  5. Emit signals. Set compat.breaking_count = number of CRITICAL-severity unversioned breaking deltas; set compat.severity_max = the highest severity observed (one of none/info/minor/major/critical).
  6. Rule. Verdict is FAIL if compat.breaking_count > 0 (any unversioned breaking change). WARN if only versioned/deprecated breaks or documented default changes exist. PASS only if every delta is additive or non-breaking. Cite the prior-release evidence for each finding (<path>:<line> old vs new); an unsupported claim of breakage is not a finding.

Read the full file on GitHub · 56 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 56 lines · 59 tokens per session scan A ebfb98a7c015

Subscribe to this mod's changes

evolve-compat-surface-check is an agent published in the GitHub repository mickeyyaya/evolve-loop (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 59 tokens to every session and 1,332 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories