VERIFICATION METHOD

Evidence is reported by level

A code example can be syntactically valid without being proven in a live colony. The site separates documentation review, offline checks, Console tests, and multi-tick room evidence instead of treating them as the same claim.

  1. 01
    Official documentation checked

    The API name, arguments, return values, limits, and timing rules are compared with current official documentation.

  2. 02
    JavaScript syntax checked

    Code is parsed or type-checked so that basic syntax and obvious type mistakes are caught before publication.

  3. 03
    Offline simulation passed

    Relevant branches are exercised with controlled mock data when a full Screeps runtime is not required.

  4. 04
    Screeps Console tested

    The example or probe is executed in the Screeps Console and the observed return value is recorded.

  5. 05
    Live multi-tick test completed

    The behavior is observed in a real room across multiple ticks, including state changes and edge conditions.

CAPTURE KIT

Capture runtime evidence in one standard format

The Console Capture Kit reads selected object state, records explicit return values, and emits a versioned JSON Evidence Bundle. It never executes movement, Spawn, Market, Link, or other game actions; multi-tick sessions only stage bounded samples in its own Memory namespace.

Open the Console Capture Kit →

COVERAGE

See what still needs runtime evidence

The coverage registry maps high-frequency symptom paths to Errors, APIs, Object Hubs, Guides, and Tools, then shows the accepted Console/live evidence level and the next evidence gap.

Review verification coverage →

VERIFIED ARCHIVE

Runtime-tested guides stay separate

Only guides whose shared source records contain explicit Console or live multi-tick evidence appear in the recently verified archive.

Browse recently verified guides →

TRANSPARENCY

Pending tests remain visible

If a guide has not been tested in the Console or in a live room, that status remains pending rather than being inferred from documentation alone.

Review the live-evidence backlog →

CONTRIBUTE EVIDENCE

Submit a reproducible observation

A useful report includes the server or shard, room state, tick range, exact code or action, return values, later-tick state, and limitations. Credentials and unrelated player data must be removed.

Open the live-evidence form ↗

SAFETY

Read-only probes come first

Diagnostics should begin with object reads, return-code logging, and state snapshots before spawning, moving, trading, destroying, or modifying Memory.

ACCEPTANCE RULE

One screenshot is not always enough

Stateful behavior normally needs a starting state, the action return code, and one or more later ticks. Accepted evidence is added with its environment and limitations, not generalized beyond what it proves.