Crate cordance_scan 

Repository scanners. Deterministic surface classification by path and filename.

Walks a target directory, hashes every file with SHA-256, and assigns each one a SourceClass taxonomy bucket (e.g. ProjectAdr, EngineeringDoctrinePrinciple, ProjectAgentFile, BlockedSurface). Classification is content-agnostic by design — it never reads prose to decide what something is, only path and filename.

Block rules cover runtime exhaust (.cordance/, .git/, .claude/cache/, node_modules/, target/, …), secret/credential filenames (id_rsa, .env, secrets.json, Credentials.*), and OS junk (.DS_Store, Thumbs.db). Case folding keeps SECRET-FOO.TXT blocked alongside secret-foo.txt on default-case-insensitive NTFS / APFS.

Golden path

use camino::Utf8PathBuf;

let target = Utf8PathBuf::from(".");
let sources = cordance_scan::scan_repo(&target).expect("scan succeeds");

for record in &sources {
    if record.blocked {
        eprintln!(
            "blocked: {} ({})",
            record.path,
            record.blocked_reason.as_deref().unwrap_or("?"),
        );
    } else {
        println!(
            "{:?}: {} ({} bytes)",
            record.class, record.path, record.size_bytes,
        );
    }
}

Modules

blocked

Blocked-surface rules. Distinguishes runtime exhaust (always blocked) from repo-tracked agent material (allowed, classified as ProjectAgentFile).

classifier

Path → SourceClass. Pure function; no I/O.

hasher

Stable sha256 of a path’s bytes.

walker

Directory walker. Honours .gitignore via the ignore crate.

Enums

ScanError

Functions

classify_by_path

Classify a path using only its repo-relative location. No content reads.

scan_repo

Top-level scan entrypoint. Delegates to walker::walk.