Documentation
¶
Overview ¶
Package coverage models codefit's coverage manifest (PRD section 10, RF-07): the explicit, per-language declaration of what codefit detects and what it does not. It turns the blind spot from "invisible and dangerous" into "declared and known", and is the single source for both the human-facing COVERAGE.md and the agent-facing codefit-coverage tool, and for the report's coverage_note.
Status: SKELETON. This declares the Manifest type. Each LanguageProvider supplies its manifest and the report derives its coverage_note from it in Fase 1; at that point the report's coverage_note stops being omitempty and is always populated (there is always something to declare about coverage).
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Manifest ¶
type Manifest struct {
Language string
// Deterministic lists the classes codefit detects with coded patterns
// (certainty 1.0): hardcoded secrets, SQL/command injection, weak crypto, ...
Deterministic []string
// Reasoning lists the classes codefit maps as surface for the agent to
// reason about: IDOR, broken authz, over-fetching, ...
Reasoning []string
// NotCovered lists, explicitly, what codefit does not audit: race
// conditions, architectural design flaws, business-logic correctness, ...
NotCovered []string
}
Manifest declares, for one language, which classes of problems codefit covers and how — and, honestly, which it does not (PRD section 10).
Source Files