@ruso-0/nreki 10.5.0 → 10.5.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +16 -0
- package/package.json +1 -1
- package/templates/CLAUDE.md +2 -1
package/CHANGELOG.md
CHANGED
|
@@ -2,6 +2,22 @@
|
|
|
2
2
|
|
|
3
3
|
All notable changes to NREKI will be documented in this file.
|
|
4
4
|
|
|
5
|
+
## 10.5.1 (2026-04-15) — Dynamic Risk Expansion
|
|
6
|
+
|
|
7
|
+
### Changed
|
|
8
|
+
- **`handleOutline` auto-expand es ahora dinámico (knapsack):** en lugar de los 3 símbolos HIGH-risk más grandes fijos, expande todos los que quepan en un presupuesto de 6,000 tokens. Nuevo umbral de tamaño por símbolo sube de 100 a 150 líneas.
|
|
9
|
+
- **Warning `[BUDGET LIMIT REACHED]`** con lista de los primeros 8 símbolos omitidos y comando `nreki_code action:"compress" focus:"..."` listo para copiar.
|
|
10
|
+
- **Orden de expansión** dentro del outline: una vez seleccionados por presupuesto, se re-ordenan por `startLine` ascendente para que el lector recorra el archivo linealmente.
|
|
11
|
+
|
|
12
|
+
### Fixed
|
|
13
|
+
- **`computeTriageRisk` filtro anti-trivialidad:** símbolos de ≤3 líneas ahora restan 2 al score. Evita marcar getters/constantes/exports triviales como HIGH-risk.
|
|
14
|
+
|
|
15
|
+
### Docs
|
|
16
|
+
- `templates/CLAUDE.md` y `skills/SKILL.md` actualizados con la nueva política de presupuesto y la instrucción crítica de usar `compress` cuando aparece `[BUDGET LIMIT REACHED]`.
|
|
17
|
+
|
|
18
|
+
### Tests
|
|
19
|
+
- 729/729 pasan (45 archivos, ~142s).
|
|
20
|
+
|
|
5
21
|
## 10.5.0 (2026-04-15) — Pre-Launch Security Hardening
|
|
6
22
|
|
|
7
23
|
### Security (Critical)
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@ruso-0/nreki",
|
|
3
|
-
"version": "10.5.
|
|
3
|
+
"version": "10.5.3",
|
|
4
4
|
"description": "MCP plugin that validates AI agent edits in RAM before they touch disk. Spectral clustering, architecture diffs, bridge detection, dead code oracle, and cross-file semantic checks for TypeScript, Go (gopls), and Python (pyright). Zero cloud dependencies.",
|
|
5
5
|
"main": "dist/index.js",
|
|
6
6
|
"type": "module",
|
package/templates/CLAUDE.md
CHANGED
|
@@ -11,7 +11,8 @@ If nreki_* tools fail: fall back to native Read/Write and inform user.
|
|
|
11
11
|
- NEVER read functions one by one. If you need 5 functions, use ONE call with comma-separated targets:
|
|
12
12
|
nreki_code action:"compress" focus:"func1, func2, func3, func4, func5"
|
|
13
13
|
- Large files (>100L): nreki_navigate action:"outline" → identify HIGH-risk methods → compress ALL in one call.
|
|
14
|
-
- The outline auto-expands
|
|
14
|
+
- The outline auto-expands HIGH-risk functions up to a 6,000 token budget. Read them directly from the outline.
|
|
15
|
+
- **CRITICAL FOR AUDITS:** If the outline says [BUDGET LIMIT REACHED], you MUST use `nreki_code action:"compress" focus:"<omitted_symbols>"` to read the remaining high-risk functions before concluding your audit. Do not guess their logic.
|
|
15
16
|
- Small files: nreki_code action:"read".
|
|
16
17
|
- Navigation: use nreki_navigate (search, definition, references, outline, map). No grep/glob.
|
|
17
18
|
|