blitzstrike 1.0.26 → 1.0.27

package/CHANGELOG.md

@@ -5,6 +5,75 @@ All notable changes to Blitz Strike are documented here.
 The format follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/).
 
+## [1.0.26] — 2026-09-13
+
+The first hardening milestone: the whole roadmap is implemented, measured, and
+type-safe. "A scan hit is a hypothesis; a live test is the verdict" is now a
+code path, not a slogan.
+
+### Added
+
+- **STRIKE validation engine** (`src/strike.ts`) — baseline + marker +
+  negative-control requests produce a deterministic verdict (`confirmed` /
+  `false_positive` / `unconfirmed` / `blocked`), wired into the finding
+  lifecycle via `resolveFinding` (`hypothesis → validating → confirmed`).
+- **`strike_verify`** rewritten and **`strike_resolve`** added — attach a live
+  verdict to a finding, advancing its lifecycle + attaching SHA-256 evidence.
+- **CVSS v3.1 calculator** (`src/cvss.ts`) + `cvss_score` tool — self-computed
+  base score, vector, and severity (verified against known NVD values), not
+  read from NVD.
+- **Finding deduplication** (`src/dedup.ts`) + `dedup_findings` tool — collapse
+  findings that share a root cause (sink × source × CWE) across engagements.
+- **Reproducible report generator** (`src/report.ts`) + `generate_report` tool —
+  deterministic markdown/JSON report with summary + SHA-256 integrity hash.
+- **Benchmark framework** (`src/benchmark.ts` + `bench/corpus/`) + `run_benchmark`
+  tool — labelled vulnerable/safe corpus across 4 languages measuring detection
+  rate, false-positive rate, and precision.
+- **Coverage matrix** (`src/coverage.ts`) + `coverage_matrix` tool — enumerates
+  language × sink-class coverage (4 languages × 12 sink classes).
+- **Dynamic guidance** (`src/guidance.ts`) — every trace/scan/verify result
+  returns `next_steps` directives so a driving agent knows the next move.
+- **Modular live reconnaissance** (`src/live-recon.ts`) + `live_recon` tool —
+  fingerprint, WAF/tech/version detection, crawler, parameter discovery,
+  subdomain enumeration, Wayback CDX, API discovery, port scan, intel correlation.
+- **Universal multi-language taint** (`src/universal-taint.ts` + `src/adapters.ts`)
+  — PHP (php-parser), JavaScript/TS (@babel/parser), Python (@lezer/python),
+  Java (java-parser), all pure-JS (no tree-sitter native addons).
+- **3 memory/confidence tools** — `memory_forget`, `confidence_weights`,
+  `finding_attach_evidence` (wired previously-dead exports into the surface).
+- **CI workflow** (`.github/workflows/ci.yml`) — typecheck + build + regression
+  suite on every push and pull request.
+- **47 missing tool manuals** generated (`scripts/generate-manuals.py`) —
+  manual coverage went 270 → 317 indexed manuals (334 files), 100% of the
+  130-tool catalog is now documented.
+
+### Changed
+
+- **MCP surface grew 33 → 54 tools.**
+- `bun run test` now runs the formal regression + benchmark suite
+  (106 checks), replacing the ad-hoc `test/verify.ts` invocation.
+
+### Fixed
+
+- **Node crash (latent)** — `require("node:fs")` inside ESM modules
+  (`memory.ts`, `benchmark.ts`) worked under Bun but threw `require is not
+  defined` under plain Node (the `npx` path). Replaced with top-level imports.
+- **`trace_data_flow` next_steps broken** — `taintedSinks` collected SinkClass
+  *objects* instead of string sink ids, so sink-specific guidance never matched.
+  Now collects `sink.id`.
+- **`strike_verify` ignored its `timeout` parameter** — now threaded into the
+  request.
+- **Dead code eliminated** — 12 unused locals/params/imports + 4 dead exports +
+  1 dead re-export, all removed or wired into tools.
+
+### Hardening
+
+- **Type safety** — 59 scattered `any` usages reduced to 0 literals. Internal
+  JSON/data types got explicit interfaces (`WafDetection`, `TechCorrelationResult`,
+  `EnrichResult`, `ScopeResult`, `NvdResponse`, `FofaResponse`, …). The only
+  remaining `any` is a single documented `AstNode` alias marking the external
+  untyped-parser boundary (php-parser / babel / lezer / java-parser).
+
 ## [1.0.0] — 2026-09-12
 
 ### Added

package/README.md

@@ -161,12 +161,14 @@ MCP servers. With no client detected, it prints the snippet for manual paste.
 
 | Tool | Purpose |
 |---|---|
-| `strike_verify(url, method, data, headers, marker, timeout)` | Live HTTP verification with marker + negative control. |
+| `strike_verify(url, method, data, headers, marker, timeout)` | Live HTTP verification with marker + negative control + baseline. |
+| `strike_resolve(finding, verdict)` | Attach a live STRIKE verdict to a finding and advance its lifecycle. |
 | `scope_check(target, scope, mode)` | Enforce scope before active testing (no-DoS, exclusion-aware, mode-gated). |
 | `run_engagement(target, scope, mode, max_files)` | Full 3-tier audit in ONE call — scope gate → triage → chain enrichment → findings. |
 | `list_chains()` | List all escalation chains in the data layer. |
 | `fofa_search(query, size, fields)` | FOFA asset index search (needs `FOFA_EMAIL` + `FOFA_KEY`). |
 | `nvd_lookup(cve_id)` | CVE lookup from NVD 2.0 (no key required). |
+| `live_recon(url)` | Modular passive-first recon: fingerprint, WAF/tech/version, crawler, params, subdomains, Wayback, API discovery, port scan. |
 
 ### CATALOG — breadth layer (tools + skills knowledge base)
 
@@ -183,11 +185,32 @@ MCP servers. With no client detected, it prints the snippet for manual paste.
 
 | Tool | Purpose |
 |---|---|
-| `read_tool_manual(name)` | Read a full deep manual for a tool (270+ manuals). |
+| `read_tool_manual(name)` | Read a full deep manual for a tool (317+ manuals). |
 | `list_manuals()` | List all manuals + playbooks. |
 | `read_playbook(name)` | Read an engagement playbook (web-app, api-security, AD, etc.). |
 | `list_playbooks()` | List all 17 engagement playbooks. |
 
+### PHASE 4 — verification, measurement, reporting
+
+| Tool | Purpose |
+|---|---|
+| `cvss_score(AV, AC, PR, UI, S, C, I, A)` | Compute a deterministic CVSS v3.1 base score + vector + severity (self-computed, not read from NVD). |
+| `dedup_findings(findings)` | Collapse findings that share a root cause (sink × source × CWE) into one group per root cause. |
+| `generate_report(findings, format)` | Emit a reproducible markdown/JSON report with summary + SHA-256 integrity hash. |
+| `run_benchmark()` | Run the labelled corpus and report detection rate, false-positive rate, and precision. |
+| `coverage_matrix()` | Enumerate language × sink-class coverage (4 languages × 12 sink classes) + coverage ratio. |
+
+### EAGLE-EYE — taint + data-flow
+
+| Tool | Purpose |
+|---|---|
+| `taint_file(path)` | Inter-procedural taint analysis (PHP) — reachable, sanitized, authorized sinks. |
+| `taint_scan(code, language)` | Universal taint analysis (PHP/JS/TS/Python/Java) with source→sink tracing. |
+| `taint_tree(path)` | Taint propagation tree — how a source flows to a sink. |
+| `trace_data_flow(path)` | Window-based data-flow trace with sanitizer + auth-gate awareness. |
+| `variant_analysis(path)` | Group reachable sinks into variant families. |
+| `list_languages()` | List supported analysis languages + extensions. |
+
 ### INTELLIGENCE — data layer (WAF + correlations + fuzzer)
 
 | Tool | Purpose |
@@ -207,7 +230,8 @@ data, vuln ontology, exploit payloads, nuclei detection templates) is sourced
 from airecon (MIT), PayloadsAllTheThings (MIT), and nuclei-templates (MIT) —
 loaded at runtime and wired into the tool surface above.
 
-270 tool manuals + 17 playbooks from kali-pentest (Apache-2.0). These are NOT
+317 indexed tool manuals + 17 playbooks from kali-pentest (Apache-2.0) — 334
+manual files, 100% of the 130-tool catalog documented. These are NOT
 decoration — they are wired into the flow:
 
 - `tool_lookup(name)` auto-attaches the tool's full manual.
@@ -220,6 +244,9 @@ decoration — they are wired into the flow:
 | `remember(topic, content, type, tags, verified)` | Save a reusable insight (deduped). verified=true only if marker reflected + negative control inert. |
 | `memory_lookup(query)` | Search memory by topic/tag/content, scored. |
 | `memory_list()` | List all memory entries, grouped by type. |
+| `memory_forget(id)` | Remove a memory entry (append-only tombstone). |
+| `confidence_weights(...)` | Get or set the confidence-scoring weight factors. |
+| `finding_attach_evidence(finding, type, description, content)` | Attach a redacted + hashed evidence record to a finding. |
 
 Memory is append-only JSONL at `~/.blitzstrike/memory.jsonl` (override with
 `BLITZSTRIKE_HOME`). `run_engagement` auto-captures matched escalation chains
@@ -305,11 +332,29 @@ All other tools need no credentials.
 bun install          # deps
 bun run typecheck    # tsc --noEmit
 bun run build        # ESM bundle → dist/index.js
+bun run test         # regression + benchmark suite (106 checks)
 bun run compile      # single static binary → dist/blitzstrike
 ```
 
 ---
 
+## Quality (measured, not claimed)
+
+Blitz Strike carries a benchmark suite and a coverage matrix so quality is a
+number you can re-run, not a claim.
+
+- **Benchmark** (`run_benchmark`) — a labelled corpus of vulnerable + safe
+  fixtures across PHP, JavaScript, Python, and Java. Current result:
+  detection rate **1.0**, false-positive rate **0**, precision **1.0**.
+- **Coverage matrix** (`coverage_matrix`) — 4 languages × 12 sink classes =
+  35/48 pairs covered (**72.9%**). The two uncovered classes (`xml_processing`,
+  `archive_extraction`) are tracked in the roadmap.
+- **Regression suite** — `bun run test` runs 106 checks covering finding
+  lifecycle, evidence integrity, CVSS math, taint tracing, dedup, report
+  integrity, and benchmark invariants. CI runs it on every push.
+
+---
+
 ## Methodology Notes
 
 - **Sink ≠ vuln.** A dangerous function in the same *file* as an unauth handler does not mean the handler calls it. Use `eagle_eye` to confirm scope.

package/ROADMAP.md

@@ -2,15 +2,28 @@
 
 Blitz Strike is version 1.0. This is where it's headed.
 
+## Done
+
+- [x] **npm publish** — `npx blitzstrike` zero-install for every MCP client.
+- [x] **Phase 1 — Finding engine** — canonical Finding schema, evidence +
+      SHA-256 integrity, secret redaction, confidence scoring, strict lifecycle.
+- [x] **Phase 2 — Source-to-sink tracing** — universal multi-language taint
+      (PHP/JS/TS/Python/Java), sanitizer + auth-gate awareness, variant grouping.
+- [x] **Phase 3 — STRIKE validation** — baseline + marker + negative control →
+      deterministic verdict, wired into the finding lifecycle.
+- [x] **Phase 4 — Benchmark & regression** — labelled corpus (detection rate /
+      false-positive rate / precision), coverage matrix, CI, formal test suite.
+- [x] **Type hardening** — 59 scattered `any` → 0 literals; explicit interfaces
+      for all internal JSON/data shapes.
+
 ## Near term
 
-- [ ] **npm publish** — `npx blitzstrike` zero-install for every MCP client.
-- [ ] **Live catalog** — `tools-catalog.json` + intelligence data fetched from
-      GitHub at startup (always current, no re-clone).
 - [ ] **Client-side pentest** — Playwright browser automation for DOM XSS,
       auth-flow replay, and CSRF PoC generation (verifiable, not decorative).
-- [ ] **`bun test` formal suite** — convert `test/verify.ts` to `bun:test` with
-      CI.
+- [ ] **XML + archive sink coverage** — `xml_processing` and `archive_extraction`
+      are the two sink classes still uncovered (see `coverage_matrix`).
+- [ ] **Java/Go/Ruby taint precision** — Java adapter is currently regex-oriented;
+      Go and Ruby are not yet covered by the universal engine.
 
 ## Medium term
 

package/docs/findings.md

@@ -89,7 +89,34 @@ validation + evidence.
 | `finding_create` | create a canonical finding (default `detected`) |
 | `finding_transition` | advance lifecycle (rejects illegal moves) |
 | `confidence_score` | deterministic weighted confidence |
+| `confidence_weights` | get/set the confidence weight factors |
 | `redact` | redact secrets from arbitrary text |
+| `finding_attach_evidence` | attach a redacted + hashed evidence record to a finding |
+| `strike_resolve` | attach a live STRIKE verdict to a finding and advance its lifecycle |
 
 `run_engagement` now emits canonical findings (status `hypothesis`, carrying
 static + sink evidence) instead of ad-hoc `{chain_id, name, severity}` tuples.
+
+## STRIKE verdict → lifecycle
+
+`src/strike.ts` runs a baseline, a marker, and a negative-control request and
+maps the result to a deterministic verdict that resolves a finding:
+
+| Verdict | Condition | Lifecycle effect |
+| :--- | :--- | :--- |
+| `confirmed` | marker reflected, control inert | `hypothesis → validating → confirmed` (+ evidence, confidence 1.0) |
+| `false_positive` | marker + control both reflected | terminal `false_positive` |
+| `unconfirmed` | marker not reflected | stays `validating` (validation `unconfirmed`) |
+| `blocked` | request failed/blocked | terminal `blocked` |
+
+## Deduplication, CVSS & reports
+
+- **Dedup** (`src/dedup.ts` + `dedup_findings`) — findings that share a root
+  cause (`sink type × source type × CWE`) collapse into one group, so the same
+  bug reached the same way in two files counts once.
+- **CVSS** (`src/cvss.ts` + `cvss_score`) — a self-computed CVSS v3.1 base score
+  + vector + severity (verified against known NVD values), instead of only
+  reading a score from NVD.
+- **Reports** (`src/report.ts` + `generate_report`) — deterministic markdown or
+  JSON with a summary table and a SHA-256 integrity hash; same findings →
+  byte-identical output.

package/docs/usage.md

@@ -45,11 +45,15 @@ Only after `strike_verify` reflects your marker AND the chain's
 
 - `tool_lookup` / `ensure_tool` — find + auto-install a security tool.
 - `skill_lookup` / `read_skill` — search 32 universal playbooks.
-- `read_tool_manual` / `read_playbook` — 270 manuals + 17 playbooks.
+- `read_tool_manual` / `read_playbook` — 317 manuals + 17 playbooks.
 - `detect_waf` / `tech_correlation` / `cve_correlation` / `port_correlation` —
   intelligence data layer lookups.
 - `payload_lookup` / `template_lookup` — exploit payloads + nuclei templates.
-- `remember` / `memory_lookup` — persist + recall verified knowledge.
+- `remember` / `memory_lookup` / `memory_forget` — persist + recall + remove
+  verified knowledge.
+- `cvss_score` / `dedup_findings` / `generate_report` — self-computed CVSS,
+  root-cause dedup, and reproducible reports.
+- `run_benchmark` / `coverage_matrix` — measure detection quality and coverage.
 
 ## Memory
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "blitzstrike",
-  "version": "1.0.26",
+  "version": "1.0.27",
   "description": "Blitz Strike — a universal MCP security-audit toolbelt. BLITZ sweeps the attack surface, EAGLE-EYE traces source-to-sink, STRIKE verifies live. 57 attack chains, 130-tool catalog, intelligence data layer. One server, every agent.",
   "type": "module",
   "bin": {