Directories
¶
| Path | Synopsis |
|---|---|
|
adapters
|
|
|
java
Package java adapts Java sources into raw API candidates — the intermediate extraction layer between the file walk (internal/discovery) and the Spring-aware IR mapping (w3-02, the architecture boundary this package must not cross).
|
Package java adapts Java sources into raw API candidates — the intermediate extraction layer between the file walk (internal/discovery) and the Spring-aware IR mapping (w3-02, the architecture boundary this package must not cross). |
|
spring
Package spring turns the raw Java candidates extracted by adapters/java (w3-01) into a meaningful Spring Boot ir.ApiSurface (w3-02): controllers become Services, mapping annotations become Methods with merged paths, binding annotations become Params/Payload, ResponseEntity/Mono/Flux returns unwrap into Response, Pageable/page-size parameters become Pagination, @ControllerAdvice handlers become the ErrorScheme, referenced records/POJOs become Types, and build files pin Source.Framework.
|
Package spring turns the raw Java candidates extracted by adapters/java (w3-01) into a meaningful Spring Boot ir.ApiSurface (w3-02): controllers become Services, mapping annotations become Methods with merged paths, binding annotations become Params/Payload, ResponseEntity/Mono/Flux returns unwrap into Response, Pageable/page-size parameters become Pagination, @ControllerAdvice handlers become the ErrorScheme, referenced records/POJOs become Types, and build files pin Source.Framework. |
|
cmd
|
|
|
vanguard
command
Command vanguard is the CLI entry point of the Vanguard linter.
|
Command vanguard is the CLI entry point of the Vanguard linter. |
|
internal
|
|
|
benchgen
Package benchgen deterministically generates the vanguard performance bench repository (w4-03; test strategy §5.1).
|
Package benchgen deterministically generates the vanguard performance bench repository (w4-03; test strategy §5.1). |
|
buildconf
Package buildconf carries the cgo link-time support files used by the release build matrix (w6-01).
|
Package buildconf carries the cgo link-time support files used by the release build matrix (w6-01). |
|
cli
Package cli implements the vanguard command-line surface (charter §6.1).
|
Package cli implements the vanguard command-line surface (charter §6.1). |
|
corpus
Package corpus guards the adversarial corpus contract (test-strategy §3): the manifest.yaml schema the w4-02 mutation harness reads, the manifest.json twin the corpus-check tool reads, and the per-case file discipline (≤ 40 lines, machine header, lure line).
|
Package corpus guards the adversarial corpus contract (test-strategy §3): the manifest.yaml schema the w4-02 mutation harness reads, the manifest.json twin the corpus-check tool reads, and the per-case file discipline (≤ 40 lines, machine header, lure line). |
|
discovery
Package discovery walks a source tree, detects language and framework, and selects the adapter that parses the API surface — the machinery behind "vanguard scan /path auto-detects everything" (charter §5.3, R1).
|
Package discovery walks a source tree, detects language and framework, and selects the adapter that parses the API surface — the machinery behind "vanguard scan /path auto-detects everything" (charter §5.3, R1). |
|
engine
Package engine implements the data-driven rule engine: selector and predicate filtering over the IR, the rule registry that binds rule metadata (YAML, go:embed) to check functions, config (.vanguard.yaml) parsing and discovery, suppression handling, and per-rule panic isolation.
|
Package engine implements the data-driven rule engine: selector and predicate filtering over the IR, the rule registry that binds rule metadata (YAML, go:embed) to check functions, config (.vanguard.yaml) parsing and discovery, suppression handling, and per-rule panic isolation. |
|
golden
Package golden is the W2 golden fixture harness (w2-07; test strategy §2.1, tier 1): it runs the REAL vanguard binary via exec — exactly the way CI invokes it — against fixture directories under testdata/golden/, snapshots the output of every format, and compares the normalized bytes against the snapshot files committed next to each fixture.
|
Package golden is the W2 golden fixture harness (w2-07; test strategy §2.1, tier 1): it runs the REAL vanguard binary via exec — exactly the way CI invokes it — against fixture directories under testdata/golden/, snapshots the output of every format, and compares the normalized bytes against the snapshot files committed next to each fixture. |
|
ir
Package ir defines the ApiSurface intermediate representation that every adapter produces and every rule consumes.
|
Package ir defines the ApiSurface intermediate representation that every adapter produces and every rule consumes. |
|
mutation
Package mutation is the W4 mutation harness (w4-02; test strategy §4): it takes the corpus's "ok" fixtures — code the linter correctly stays silent on — applies a mutator that turns them into real violations, and demands the target rule fire.
|
Package mutation is the W4 mutation harness (w4-02; test strategy §4): it takes the corpus's "ok" fixtures — code the linter correctly stays silent on — applies a mutator that turns them into real violations, and demands the target rule fire. |
|
perf
Package perf hosts the w4-03 tier-4 harness tests (test strategy §2.4): determinism of the real binary's output across repeated runs (§5.4) and the edge-case behaviour contract (§5.5).
|
Package perf hosts the w4-03 tier-4 harness tests (test strategy §2.4): determinism of the real binary's output across repeated runs (§5.4) and the edge-case behaviour contract (§5.5). |
|
render
Package render renders findings in the three output formats promised by the charter: pretty (human, grouped by rule), JSON (schema 1) and SARIF 2.1.0 (charter §6.6–6.8; implementation w2-05).
|
Package render renders findings in the three output formats promised by the charter: pretty (human, grouped by rule), JSON (schema 1) and SARIF 2.1.0 (charter §6.6–6.8; implementation w2-05). |
|
robustness
Package robustness hosts the W4-05 hostile-input tests (GATE CH3, brief deliverable 1): the scanner's behaviour against malicious inputs, as opposed to the merely extreme ones of w4-03's edge cases (test-strategy §5.5 draws exactly this boundary).
|
Package robustness hosts the W4-05 hostile-input tests (GATE CH3, brief deliverable 1): the scanner's behaviour against malicious inputs, as opposed to the merely extreme ones of w4-03's edge cases (test-strategy §5.5 draws exactly this boundary). |
|
Package rules hosts the R1xx–R6xx rule families: one metadata record per rule (id, slug, category, default severity, AIP reference, good/bad examples) stored as data and embedded into the binary at build time, plus the Go check functions and the registry that binds them (charter §3 and §5.1; implementation: w3-03..w3-06).
|
Package rules hosts the R1xx–R6xx rule families: one metadata record per rule (id, slug, category, default severity, AIP reference, good/bad examples) stored as data and embedded into the binary at build time, plus the Go check functions and the registry that binds them (charter §3 and §5.1; implementation: w3-03..w3-06). |
|
tools
|
|
|
audit-check
command
audit-check — the FP/FN audit harness (test-strategy §7, task w4-04).
|
audit-check — the FP/FN audit harness (test-strategy §7, task w4-04). |
|
corpus-check
command
corpus-check runs the adversarial corpus manifest (testdata/adversarial, test-strategy §3.6) against a vanguard binary and reports agreement.
|
corpus-check runs the adversarial corpus manifest (testdata/adversarial, test-strategy §3.6) against a vanguard binary and reports agreement. |
|
gen-bench-repo
command
Command gen-bench-repo generates the vanguard performance bench repository (w4-03; test strategy §5.1).
|
Command gen-bench-repo generates the vanguard performance bench repository (w4-03; test strategy §5.1). |
|
perfbench
command
Command perfbench drives the w4-03 performance + determinism bench (test strategy §5) against a sandbox build of vanguard.
|
Command perfbench drives the w4-03 performance + determinism bench (test strategy §5) against a sandbox build of vanguard. |
|
qa-label
command
qa-label applies the w4-04 DRY-RUN provisional verdicts to the machine sheet — PIPELINE QA ONLY (protocol §0: the gate labels the real sheet; this artifact is evidence that the merge path works end-to-end and is never presented as the audit result).
|
qa-label applies the w4-04 DRY-RUN provisional verdicts to the machine sheet — PIPELINE QA ONLY (protocol §0: the gate labels the real sheet; this artifact is evidence that the merge path works end-to-end and is never presented as the audit result). |
|
sarif-validate
command
sarif-validate checks that SARIF documents are valid SARIF 2.1.0 — the GATE CH3 re-validation of vanguard's SARIF output (w4-05 deliverable 2) over the adversarial corpus and the military-youth audit sample.
|
sarif-validate checks that SARIF documents are valid SARIF 2.1.0 — the GATE CH3 re-validation of vanguard's SARIF output (w4-05 deliverable 2) over the adversarial corpus and the military-youth audit sample. |
Click to show internal directories.
Click to hide internal directories.