config-audit/scanners/lib/humanizer.mjs
Kjell Tore Guttormsen 97867dbf37 fix(acr): AGT findings humanize to "Wasted tokens" not "Other" (M-BUG-17)
The agent-listing scanner (AGT) emits an always-loaded per-turn token cost
("Agent description is long, re-sent every turn in the always-loaded listing";
scanner category 'token-efficiency', "the dominant single always-loaded
source"). But SCANNER_TO_CATEGORY in humanizer.mjs had no AGT entry, so its
findings fell through to the 'Other' fallback (humanizer.mjs:140) — a bucket
that isn't even in the analyzer-agent's category list. Neither the scanner
prefix nor the per-finding category ('token-efficiency' is not in
CATEGORY_TO_IMPACT) resolved AGT to its true impact.

Same class as M-BUG-16/15: a finding type without its matching humanizer
mapping landing on a default that mismatches its own evidence. The analogous
SKL body finding correctly buckets "Wasted tokens"; AGT (the same always-loaded
token-waste mechanism) silently landed under the meaningless "Other".

Found during analyze-prep premise-verification of the linkedin-posts scan: the
3 AGT findings bucketed "Other" while the analogous SKL findings bucketed
"Wasted tokens".

Fix: add AGT: 'Wasted tokens' to SCANNER_TO_CATEGORY, alongside TOK/CPS/SKL.
RED-first (extended the Wasted-tokens category test to include AGT; the 'Other'
fallback test still uses a synthetic 'XXX' scanner, unaffected). Frozen v5.0.0
untouched (AGT post-dates it; humanizer bypassed for --raw/--json); no
default-output snapshot contains AGT -> 0 regen. Suite 1359/0.

Verified end-to-end on linkedin-posts: 3 AGT findings now "Wasted tokens", 0
"Other" remaining, 28 findings unchanged (category-only). All 16 orchestrator
scanner prefixes now covered by the category map (class closed).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01683eAqVecv9VZfQzL8CQ9h
2026-06-30 13:51:55 +02:00

212 lines
6.7 KiB
JavaScript

/**
* Plain-language humanizer for config-audit findings.
*
* Pure functions. Never mutate inputs. Translates technical scanner output
* into user-friendly language at output-formatting time. Adds three new
* fields to each finding:
* - userImpactCategory: human-readable label per scanner (research/02)
* - userActionLanguage: one-line urgency phrase per severity
* - relevanceContext: deterministic file-pattern heuristic
*
* Original id, scanner, severity, file, line, evidence, category, autoFixable
* are preserved exactly. Title, description, recommendation are replaced when
* a translation is found; otherwise the originals are kept.
*
* Lookup order (per scanner prefix):
* 1. exact title in TRANSLATIONS[prefix].static
* 2. first regex match in TRANSLATIONS[prefix].patterns
* 3. TRANSLATIONS[prefix]._default
* 4. fallthrough: original strings (when scanner prefix has no entry)
*
* Zero external dependencies.
*/
import { TRANSLATIONS } from './humanizer-data.mjs';
/**
* Map scanner prefix to user-facing impact-category label (research/02 line 124).
*/
const SCANNER_TO_CATEGORY = {
CML: 'Configuration mistake',
SET: 'Configuration mistake',
HKV: 'Configuration mistake',
RUL: 'Configuration mistake',
MCP: 'Configuration mistake',
IMP: 'Configuration mistake',
CNF: 'Conflict',
COL: 'Conflict',
TOK: 'Wasted tokens',
CPS: 'Wasted tokens',
SKL: 'Wasted tokens',
AGT: 'Wasted tokens',
DIS: 'Dead config',
GAP: 'Missed opportunity',
PLH: 'Configuration mistake',
OST: 'Configuration mistake',
OPT: 'Missed opportunity',
};
/**
* Per-finding `category` values that override the scanner-default impact label.
* Needed when one finding inside a scanner means something different from the
* scanner's usual bucket — e.g. stale plugin-cache versions are emitted by TOK
* (normally "Wasted tokens") but load on ZERO turns, so their honest impact is
* "Dead config" (present on disk, never loaded), not wasted per-turn tokens.
*/
const CATEGORY_TO_IMPACT = {
'plugin-cache-hygiene': 'Dead config',
};
/**
* Map severity to one-line action-language phrase (research/02 line 134).
*/
const SEVERITY_TO_ACTION = {
critical: 'Fix this now',
high: 'Fix soon',
medium: 'Fix when convenient',
low: 'Optional cleanup',
info: 'FYI',
};
/**
* Compute relevance context from a finding's file path. Deterministic, in-process,
* no subprocess. Conservative — defaults to 'affects-everyone' when ambiguous.
*
* @param {string|null|undefined} filePath
* @returns {'test-fixture-no-impact' | 'affects-this-machine-only' | 'affects-everyone'}
*/
export function computeRelevanceContext(filePath) {
if (typeof filePath !== 'string' || filePath.length === 0) {
return 'affects-everyone';
}
if (filePath.includes('/tests/fixtures/') || filePath.includes('/test/fixtures/')) {
return 'test-fixture-no-impact';
}
// Match basename pattern *.local.* (e.g., settings.local.json, claude.local.md)
const basename = filePath.split('/').pop() || '';
if (/\.local\./.test(basename)) {
return 'affects-this-machine-only';
}
return 'affects-everyone';
}
/**
* Look up translation for a finding by scanner prefix and title.
* Returns the translation object or null when no match (caller falls through to original).
*
* @param {string} scanner
* @param {string} title
* @returns {{title:string, description:string, recommendation:string} | null}
*/
function lookupTranslation(scanner, title) {
const entry = TRANSLATIONS[scanner];
if (!entry) return null;
// 1. Exact static match
if (typeof title === 'string' && entry.static && Object.prototype.hasOwnProperty.call(entry.static, title)) {
return entry.static[title];
}
// 2. Pattern match
if (Array.isArray(entry.patterns) && typeof title === 'string') {
for (const p of entry.patterns) {
if (p.regex instanceof RegExp && p.regex.test(title)) {
return p.translation;
}
}
}
// 3. Default
if (entry._default) {
return entry._default;
}
return null;
}
/**
* Humanize a single finding. Pure — never mutates input. Returns a new object.
*
* @param {object} finding - finding object from scanner output
* @returns {object} new finding with translated title/description/recommendation +
* userImpactCategory, userActionLanguage, relevanceContext fields
*/
export function humanizeFinding(finding) {
if (!finding || typeof finding !== 'object') {
return finding;
}
const translation = lookupTranslation(finding.scanner, finding.title);
const category =
CATEGORY_TO_IMPACT[finding.category] || SCANNER_TO_CATEGORY[finding.scanner] || 'Other';
const action = SEVERITY_TO_ACTION[finding.severity] || 'FYI';
const relevance = computeRelevanceContext(finding.file);
const out = {
// Preserve identifying / structural fields exactly
id: finding.id,
scanner: finding.scanner,
severity: finding.severity,
// Replace prose if a translation exists; otherwise keep originals
title: translation ? translation.title : finding.title,
description: translation ? translation.description : finding.description,
file: finding.file ?? null,
line: finding.line ?? null,
evidence: finding.evidence ?? null,
category: finding.category ?? null,
recommendation: translation ? translation.recommendation : finding.recommendation,
autoFixable: finding.autoFixable ?? false,
// New humanized fields
userImpactCategory: category,
userActionLanguage: action,
relevanceContext: relevance,
};
// Preserve optional details payload if present (v5 N6)
if (finding.details && typeof finding.details === 'object') {
out.details = finding.details;
}
return out;
}
/**
* Humanize an array of findings. Pure — returns a new array of new objects.
*
* @param {object[]} findings
* @returns {object[]}
*/
export function humanizeFindings(findings) {
if (!Array.isArray(findings)) return findings;
return findings.map(humanizeFinding);
}
/**
* Humanize a top-level envelope produced by `runAllScanners`. Walks
* `env.scanners[].findings`. Pure — returns a new envelope with new
* scanner objects and new finding objects. The envelope-level shape
* (scanners array, target_path, total_duration_ms, aggregate, etc.)
* is preserved.
*
* @param {object} env
* @returns {object}
*/
export function humanizeEnvelope(env) {
if (!env || typeof env !== 'object' || !Array.isArray(env.scanners)) {
return env;
}
const newScanners = env.scanners.map((s) => {
if (!s || typeof s !== 'object') return s;
if (!Array.isArray(s.findings)) return s;
return {
...s,
findings: humanizeFindings(s.findings),
};
});
return {
...env,
scanners: newScanners,
};
}