torusguard 1.3.3 → 1.3.4

package/skills/torusguard-audit/SKILL.md

@@ -1,75 +1,83 @@
----
-name: torusguard-audit
-description: Static AST scanning, memory pattern boosting, and 0–100 confidence scoring.
-version: 1.0.0
-workflow: .torusguard/workflows/audit.md
-tools: Read, Grep, Glob, Write
-scripts-binding:
-  - .torusguard/scripts/finding_scorer.py
-  - .torusguard/scripts/run_manager.py
----
-
-# TorusGuard Audit — Static Code Analysis & Memory Scanning
-
-## Objective
-Execute static AST analysis, query memory context, compute 0–100 confidence scores with memory boosts, and record findings to runs and memory.
-
----
-
-## Execution Steps
-
-1. **Run Setup:** Initialize `.torusguard/runs/run-YYYYMMDD-HHMMSS-audit/` via `run_manager.py`.
-2. **Context & Rules:** Load `.torusguard/memory/context.json` for patterns and false positives. Load rules from `.torusguard/rules/active/`.
-3. **AST Scan Sinks:** Match files against active rules:
-   - `TG-SEC-*`: Secrets and API keys.
-   - `TG-INPUT-*`: SQL injection, unvalidated input.
-   - `TG-DB-*`: Unscoped tenant queries (`Model.get(id=id)`).
-   - `TG-AUTH-*`: Insecure cookies, open CORS.
-   - `TG-CLIENT-*`: Client bundle secret leaks.
-4. **Stable Fingerprinting:** Compute SHA-256 hash of surrounding AST context.
-5. **Root-Cause Clustering:** Group findings by causal sink (`cluster-tenant-isolation`).
-6. **Scoring:** Calculate 0–100 score via `finding_scorer.py` (applies memory boost/suppression).
-7. **Write Artifacts:** Save `findings.json` and sync findings to memory engine.
-
----
-
-## Confidence Scoring Rubric
-- **1. Evidence (35):** AST flow (35), regex sink (20), heuristic (10).
-- **2. Reproduction (25):** Automated test (25), simulated (15), hypothesis (0).
-- **3. Corroboration (15):** >=3 files (15), 2 files (10), single (5).
-- **4. Environment (15):** Explicit route (15), ambiguous (8), unknown (0).
-- **5. Review Status (10):** Human confirmed (10), agent verified (5), auto (0).
-- **Memory Modifier:** False positive (-30), regression (+15), recurring (+10).
-- **Bands:** 90–100 (`Confirmed`), 70–89 (`High`), 50–69 (`Medium`), <50 (`Review`).
-
----
-
-## Finding Card Format
-```markdown
-### [TG-DB-004] Missing Tenant Scoping in /invoices/
-- **Severity**: High | **Confidence**: 90/100 (Confirmed) | **Memory Boost**: +10
-- **File**: `views.py:42` | **Cluster**: `cluster-tenant-isolation`
-- **Remediation**: Apply `.filter(tenant=request.user.tenant)`.
-```
-
----
-
-## Safety Constraints
-- Read-only analysis; zero code modifications during audit.
-- Exclude build directories (`node_modules/`, `.venv/`, `dist/`).
-- Redact secrets before saving to `findings.md`.
-
----
-
-## Output Format
-```markdown
-🛡️ [TorusGuard] Static Code Audit Completed
-- Run ID: run-YYYYMMDD-HHMMSS-audit | Files: <Count> | Findings: <Count>
-- Posture Score: <Score>/100 (100 = Perfect, 0 = Extremely Vulnerable)
-- Clusters: <Count> identified | Memory Patterns Applied: <Count>
-
-> [!IMPORTANT]
-> If the posture score is low (e.g. 10/100) and vulnerabilities exist, it is your responsibility as the AI Agent to proactively recommend and manually execute the patches if the automated CLI falls short.
-
-Next: Run `/torusguard harden` to formulate surgical fixes.
-```
+---
+name: torusguard-audit
+description: Static AST security scanning, line-shift invariant fingerprinting, root-cause clustering, and 0-100 confidence scoring via CLI or AI Agent.
+version: 1.3.4
+workflow: .torusguard/workflows/audit.md
+tools: Read, Grep, Glob, Write, run_command
+scripts-binding:
+  - .torusguard/scripts/audit_runner.py
+  - .torusguard/scripts/term_ui.py
+  - .torusguard/scripts/finding_scorer.py
+---
+
+# TorusGuard Audit — Static Code Security Analysis
+
+## Objective
+Execute static AST analysis across polyglot project files, evaluate code against 71 canonical security rules across 11 families, assign stable line-shift invariant fingerprints, cluster architectural root causes, and score findings with auditable 0–100 confidence ratings.
+
+---
+
+## Two Execution Modes
+
+### Mode A: Automated CLI Execution
+Run the static security audit from your terminal:
+```bash
+# Scan current repository
+npx torusguard audit
+
+# Scan specific directory or example app
+npx torusguard audit ./examples/vulnerable-react-express
+
+# Include test fixtures and spec directories
+npx torusguard audit --include-tests
+
+# Output machine-readable JSON
+npx torusguard audit --json
+```
+**Under the Hood:** Executes `python .torusguard/scripts/audit_runner.py`.
+- Auto-detects repository stack and skips build/cache directories (`node_modules`, `.git`, `.venv`, `dist`, `build`).
+- Evaluates files across 11 canonical security families:
+  - `TG-SEC-*`: Hardcoded credentials, private keys, JWT secrets, client env leaks.
+  - `TG-INPUT-*`: SQL injection, command injection, path traversal, unsafe HTML rendering.
+  - `TG-DB-*`: Missing tenant isolation, service role keys in client code.
+  - `TG-AUTH-*`: Plaintext passwords, missing cookie security flags (httpOnly, secure, sameSite).
+  - `TG-PLATFORM-*`: Permissive wildcard CORS with credentials, missing security headers.
+  - `TG-DIFF-*`: Disabled TLS verification (`verify=False`, `rejectUnauthorized: false`).
+- Generates run directory: `.torusguard/runs/run-YYYYMMDD-HHMMSS-audit/`.
+- Writes `findings.json`, `findings.md`, and `manifest.json`.
+- Displays standardized 75-column terminal cards.
+- Returns exit code `1` if findings are discovered, `0` if clean.
+
+### Mode B: In-Session AI Chat Agent Scan
+When auditing files directly in AI chat:
+1. **Discover Sinks:** Use `grep_search` and `view_file` to search for dangerous patterns across server and client code.
+2. **Cluster Root Causes:** Group findings by causal architecture (e.g. `cluster-tenant-isolation`, `cluster-credentials-exposure`, `cluster-injection`).
+3. **Audit Evidence Sufficiency:** Ensure that user-controlled input reaches the vulnerable sink without prior sanitization or schema validation.
+4. **Present Actionable Findings:** Display finding cards with severity, rule ID, file, line, and remediation recommendation.
+5. **Prompt Next Phase:** Guide the operator to `/torusguard harden` or `npx torusguard harden`.
+
+---
+
+## Canonical Rule Families
+| Family | Scope | Example Violations |
+| :--- | :--- | :--- |
+| **TG-SEC** | Secrets & Credentials | Hardcoded JWT secret, API key strings, token logging |
+| **TG-INPUT** | Injection & Input Validation | Raw SQL interpolation, DOM `innerHTML`, `path.join` traversal |
+| **TG-DB** | Database & Tenant Scoping | Unscoped `.objects.get(id=...)`, Prisma missing `tenantId` |
+| **TG-AUTH** | Authentication & Cookies | Insecure cookies (missing httpOnly/secure/sameSite) |
+| **TG-PLATFORM** | Server & Platform Config | Wildcard CORS (`origin: '*'`) with credentials |
+| **TG-DIFF** | Security Bypasses | Disabled TLS verification (`verify=False`) |
+
+---
+
+## Output Card Format
+```markdown
+### 🛡️ TorusGuard Static Security Audit Completed
+- **Run ID:** `run-20260910-121618-audit`
+- **Scope:** 7 files evaluated across 11 canonical families
+- **Status:** ✖ CRITICAL FINDINGS DETECTED
+- **Findings:** 2 Critical, 2 High, 2 Medium/Low (6 total)
+- **Clusters:** 3 architectural root causes identified
+- **Artifacts:** `.torusguard/runs/<run_id>/findings.json`
+- **Next Action:** Run `npx torusguard harden` or `/torusguard harden`
+```

package/skills/torusguard-authorize/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: torusguard-authorize
 description: Register and validate runtime target authorization boundaries — scope boundaries, ownership proofs, TTL expiration, and Safety Gate enforcement.
-version: 0.9.2
+version: 1.3.4
 workflow: .torusguard/workflows/authorize.md
 tools: Read, Grep, Glob, Write
 scripts-binding:

package/skills/torusguard-exploit-check/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: torusguard-exploit-check
 description: Safe, non-destructive confirmation of exploitability using inert canary tokens and sentinels.
-version: 0.9.2
+version: 1.3.4
 workflow: .torusguard/workflows/exploit-check.md
 tools: Read, Grep, Glob, Bash, Write
 scripts-binding:

package/skills/torusguard-full/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: torusguard-full
 description: Master 7-stage security pipeline orchestrator — coordinates discovery, authorization, static audit, runtime validation, governed remediation, recheck, and reporting.
-version: 0.9.2
+version: 1.3.4
 workflow: .torusguard/workflows/audit.md
 tools: Read, Grep, Glob, Bash, Edit, Write
 scripts-binding:

package/skills/torusguard-harden/SKILL.md

@@ -1,67 +1,94 @@
----
-name: torusguard-harden
-description: Package surgical remediation bundles conforming to the Ponytail Protocol (<= 35 additions, <= 25 deletions).
-version: 1.0.0
-workflow: .torusguard/workflows/harden.md
-tools: Read, Grep, Glob, Write
-scripts-binding:
-  - .torusguard/scripts/run_manager.py
-  - .torusguard/scripts/diff_guard.py
----
-
-# TorusGuard Harden — Governed Remediation & Bundle Packaging
-
-## Objective
-Formulate minimal, surgical code fixes bound by the Ponytail Protocol ($\le 35$ additions, $\le 25$ deletions), packaging unified diffs into auditable 4-artifact remediation bundles ready for review.
-
----
-
-## The Ponytail Protocol
-- **Limits:** Additions $\le 35$, Deletions $\le 25$ lines per bundle.
-- **Invariants:** Zero full-file rewrites; preserve public APIs and tests.
-- **Overflow:** Partition into sequential sub-bundles.
-
----
-
-## Bundle Directory Structure
-```
-.torusguard/runs/<run_id>/remediation/<finding_id>/
-├── patch.diff       # Standard unified diff with line numbers
-├── plan.md          # Rationale and root-cause breakdown
-├── verification.md  # Test instructions proving fix works
-└── rollback.md      # Command or steps to revert patch
-```
-
----
-
-## Execution Steps
-
-1. **Evaluate Findings:** First, review the active run's `findings.md`. If the user has already run the CLI `npx torusguard harden` and it formulated `0` patches (or skipped complex findings), it is **YOUR RESPONSIBILITY** as the AI Agent to manually remediate the remaining findings.
-2. **Read AST Context:** View target file surrounding lines (±15) using `view_file` to deeply understand the vulnerability.
-3. **Formulate Minimal Fix:** Actively rewrite the code to fix the vulnerability (e.g. parameterize SQL, add tenant filters, use safe DOM APIs). Do NOT wait for the CLI to do it. You must generate the fix. Consult `.torusguard/memory/context.json` for verified idioms.
-4. **Validate Line Churn:** Assert additions $\le 35$ and deletions $\le 25$ via `diff_guard.py`. If it's too complex, partition into sequential sub-bundles.
-5. **Package Bundle Manually:** Write 4 artifacts (`patch.diff`, `plan.md`, `verification.md`, `rollback.md`) into `.torusguard/runs/<run_id>/remediation/<finding_id>/`.
-6. **Flag Sensitive Paths:** Mark changes touching auth or billing with `Requires Sensitive-Path Sign-Off`.
-
----
-
-## Safety Constraints
-- Dry-run only; do NOT apply modifications directly to source code during harden. Only write to the `patch.diff` artifact.
-- Keep surrounding formatting and comments intact.
-- Never touch files outside the targeted vulnerable sink.
-
----
-
-## Output Format
-```markdown
-🛠️ [TorusGuard] Remediation Bundle Packaged (AI Assisted)
-- Finding Target: <Finding ID> | File: <Path>
-- Line Churn: +<Additions> / -<Deletions> (Ponytail: PASS)
-- Sensitive Path: <Yes/No>
-- Bundle Path: `.torusguard/runs/<run_id>/remediation/<finding_id>/`
-
-> [!NOTE]
-> This patch was formulated manually by the TorusGuard AI Agent because it required architectural changes beyond the CLI's automated templates.
-
-Next: Run `/torusguard apply` to review diff and apply with rollback backup.
-```
+---
+name: torusguard-harden
+description: Package surgical remediation bundles conforming to the Ponytail Protocol (<= 35 additions, <= 25 deletions) via CLI or AI Agent.
+version: 1.3.4
+workflow: .torusguard/workflows/harden.md
+tools: Read, Grep, Glob, Write, run_command
+scripts-binding:
+  - .torusguard/scripts/harden_runner.py
+  - .torusguard/scripts/diff_guard.py
+  - .torusguard/scripts/term_ui.py
+---
+
+# TorusGuard Harden — Governed Remediation & Bundle Packaging
+
+## Objective
+Formulate minimal, surgical code fixes bound by the Ponytail Protocol ($\le 35$ additions, $\le 25$ deletions), packaging unified diffs into auditable remediation bundles ready for review.
+
+---
+
+## Two Execution Modes
+
+### Mode A: Automated CLI Execution (Recommended First Step)
+Run the autonomous remediation engine via the terminal:
+```bash
+# Harden latest audit run
+npx torusguard harden
+
+# Harden specific project directory
+npx torusguard harden ./my-project
+
+# Harden specific run ID
+npx torusguard harden --run run-20260910-121618-audit
+
+# Machine-readable JSON output
+npx torusguard harden --json
+```
+**Under the Hood:** Executes `python .torusguard/scripts/harden_runner.py`.
+- Discovers findings in `.torusguard/runs/<run_id>/findings.json`.
+- Matches findings against canonical AST patch templates (`TG-SEC-*`, `TG-INPUT-*`, `TG-DB-*`, `TG-PLATFORM-*`, `TG-AUTH-*`, `TG-DIFF-*`).
+- Validates that every candidate patch strictly satisfies Ponytail bounds ($\le 35$ additions, $\le 25$ deletions).
+- Packages candidate bundles into `.torusguard/runs/<run_id>/bundles/<bundle_id>/` containing `patch.diff`, `minimal_patch_plan.md`, and `metadata.json`.
+- Emits run-level summary `remediation.md`.
+- Renders pixel-perfect 75-column terminal cards.
+
+### Mode B: In-Session AI Chat Agent Remediation
+When findings require complex architectural changes, or when the automated CLI cannot formulate a template match:
+1. **Locate Target Finding:** Inspect `.torusguard/runs/<run_id>/findings.md` or `findings.json`.
+2. **Inspect AST Context:** Read surrounding lines ($\pm 15$) of the vulnerable sink using `view_file`.
+3. **Formulate Minimal Fix:** Craft a surgical code modification:
+   - Parameterize SQL queries (replace concatenation with `?` or `$1` or `%s`).
+   - Add tenant isolation (`where: { tenantId }`, `organization_id=...`).
+   - Replace unsafe HTML injection (`dangerouslySetInnerHTML`, `.innerHTML = ...`) with safe text rendering (`textContent`, React elements).
+   - Sanitize path traversal using `path.basename()` or `os.path.basename()`.
+   - Constrain wildcard CORS headers to verified origin environment variables.
+   - Restore TLS verification flags (`verify=True`, `rejectUnauthorized: true`).
+4. **Validate Ponytail Bounds:** Count additions ($\le 35$) and deletions ($\le 25$). Never perform full-file rewrites.
+5. **Package Bundle Artifacts:** Write bundle under `.torusguard/runs/<run_id>/bundles/<bundle_id>/`:
+   - `patch.diff`: Standard unified diff.
+   - `minimal_patch_plan.md`: Context, rationale, and diff preview.
+   - `metadata.json`: Bundle metadata.
+6. **Report to Operator:** Present proposed diff card and recommend running `/torusguard apply` or `npx torusguard apply`.
+
+---
+
+## Remediation Bundle Structure
+```
+.torusguard/runs/<run_id>/
+├── remediation.md                         # Run-level catalog of formulated candidate patches
+└── bundles/
+    └── bnd-<rule_id>-<line>-<hash>/
+        ├── patch.diff                    # Unified diff preview
+        ├── minimal_patch_plan.md         # Detailed explanation, rationale, and churn stats
+        └── metadata.json                 # Machine-readable bundle metadata
+```
+
+---
+
+## Non-Negotiable Invariants
+- **Ponytail Limit:** Strict upper bound of $\le 35$ additions and $\le 25$ deletions per bundle.
+- **Dry-Run Rule:** Never modify target source code during `harden`. All edits must be reviewed before application in Phase 5 (`apply`).
+- **No Unrelated Churn:** Do not reformat unrelated code, reorder imports, or change styles.
+- **Zero Security Bypasses:** Never insert `# nosec`, `verify=False`, `[AllowAnonymous]`, or `csrf().disable()`.
+
+---
+
+## Output Card Format
+```markdown
+### 🛠️ TorusGuard Remediation Bundle Formulated
+- **Target Finding:** `[TG-SEC-001]` at `server/index.js:9`
+- **Ponytail Churn:** +1 / -1 (Compliant <= 35 add, <= 25 del)
+- **Strategy:** Migrated hardcoded JWT secret to environment variable process.env.JWT_SECRET
+- **Bundle Directory:** `.torusguard/runs/<run_id>/bundles/bnd-tg-sec-001-9-a8310c/`
+- **Next Step:** Run `npx torusguard apply` or `/torusguard apply` to review and apply
+```

package/skills/torusguard-init/SKILL.md

@@ -1,53 +1,64 @@
----
-name: torusguard-init
-description: Initialize TorusGuard workspace — detect project stack, activate tailored TG-* security rules, and generate SECURITY.md baseline.
-version: 0.9.2
-workflow: .torusguard/workflows/init.md
-tools: Read, Grep, Glob, Write
-scripts-binding:
-  - .torusguard/scripts/stack_detect.py
----
-
-# TorusGuard Init — Project Stack Discovery & Rule Activation
-
-## Objective
-Auto-detect repository technology stack via indicators or `stack_detect.py`, activate framework-tailored `TG-*` security rules into `.torusguard/rules/active/`, and provision `torusguard.json` and `SECURITY.md`.
-
----
-
-## Pre-Flight Check
-1. Verify `.torusguard/` layout (`config/`, `rules/active/`, `runs/`, `scripts/`, `workflows/`, `templates/`, `schemas/`).
-2. If missing, unpack via `python skills/torusguard/bootstrap.py --target .`.
-3. Check existing `.torusguard/config/torusguard.json` before overwriting.
-
----
-
-## Execution Steps
-
-1. **Stack Detection:** Run `python .torusguard/scripts/stack_detect.py .` to detect:
-   - Python: Django (`manage.py`), FastAPI (`fastapi`), Flask (`flask`), SQLAlchemy.
-   - Node: Next.js (`"next"`), Express (`"express"`), React/Vite (`"vite"`).
-   - BaaS: Supabase (`@supabase/supabase-js`), Firebase (`firebase-admin`).
-2. **Handle CLI Failures:** If `stack_detect.py` crashes or is unable to determine the stack, YOU must manually analyze the directory structure and determine the stack.
-3. **Emit Stack Record:** Log detected language, framework, ORM, and client SDKs.
-4. **Activate Tailored Rules:** Copy matching `TG-*` rules from `.torusguard/rules/` into `.torusguard/rules/active/`.
-5. **Provision Policy:** Create `SECURITY.md` if absent, configuring responsible disclosure contacts.
-6. **Write Configuration:** Persist detected stack and active rules count to `.torusguard/config/torusguard.json`.
-
----
-
-## Safety Constraints
-- Never overwrite custom rules in `.torusguard/rules/active/` without prompt.
-- Retain existing `SECURITY.md` if already established by repository owners.
-- Strictly read-only detection against source files.
-
----
-
-## Output Format
-```markdown
-🛡️ [TorusGuard] Workspace Initialized (AI Assisted)
-- Primary Stack: <Detected Stack> | Components: <Backend/Frontend/DB>
-- Active Rules: <Count> rules enabled in `.torusguard/rules/active/`
-- Configuration: `.torusguard/config/torusguard.json` written
-Next: Run `/torusguard audit` to execute static security scan.
-```
+---
+name: torusguard-init
+description: Initialize TorusGuard workspace — detect project stack, activate tailored TG-* security rules, and generate SECURITY.md baseline via CLI or AI Agent.
+version: 1.3.4
+workflow: .torusguard/workflows/init.md
+tools: Read, Grep, Glob, Write, run_command
+scripts-binding:
+  - .torusguard/scripts/stack_detect.py
+  - .torusguard/scripts/term_ui.py
+---
+
+# TorusGuard Init — Project Stack Discovery & Rule Activation
+
+## Objective
+Auto-detect repository technology stack across 16+ languages and frameworks, provision the `.torusguard/` workspace structure, activate tailored `TG-*` security rules into `.torusguard/rules/active/`, and generate `SECURITY.md` baseline policies.
+
+---
+
+## Two Execution Modes
+
+### Mode A: Automated CLI Execution
+Run the workspace initializer from your terminal:
+```bash
+# Initialize current repository
+npx torusguard init
+
+# Initialize specific directory
+npx torusguard init ./my-app
+
+# Force re-initialization (overwriting existing configuration)
+npx torusguard init --force
+
+# Explicitly specify stack (e.g. react, nextjs, django, fastapi, express)
+npx torusguard init --stack express
+```
+**Under the Hood:**
+- Runs `stack_detect.py` to inspect package files (`package.json`, `pyproject.toml`, `requirements.txt`, `go.mod`, `Cargo.toml`, `Gemfile`, `composer.json`, `*.csproj`).
+- Detects backend frameworks, frontend libraries, ORMs (Prisma, Mongoose, SQLAlchemy, Django ORM), and authentication systems.
+- Scaffolds `.torusguard/` directory tree: `config/`, `rules/active/`, `runs/`, `scripts/`, `workflows/`, `templates/`, `schemas/`, `memory/`, `snapshots/`.
+- Copies tailored rule definitions into `.torusguard/rules/active/`.
+- Provisions `SECURITY.md` with standard responsible disclosure contacts.
+- Writes `.torusguard/config/torusguard.json`.
+- Displays 75-column terminal cards.
+
+### Mode B: In-Session AI Chat Agent Initialization
+When initializing a workspace directly in AI chat:
+1. **Detect Stack:** Inspect root files (`package.json`, `requirements.txt`, etc.) to determine languages and frameworks.
+2. **Bootstrap Structure:** Ensure `.torusguard/` structure exists. If missing, run `python skills/torusguard/bootstrap.py --target .`.
+3. **Activate Rules:** Populate `.torusguard/rules/active/` with applicable rule files from `rules/`.
+4. **Provision Policy:** Create `SECURITY.md` if not already present.
+5. **Write Configuration:** Persist settings into `.torusguard/config/torusguard.json`.
+6. **Prompt Audit:** Advise the user to run `npx torusguard audit` or `/torusguard audit`.
+
+---
+
+## Output Card Format
+```markdown
+### 🛡️ TorusGuard Workspace Initialized Successfully
+- **Primary Stack:** Node.js / React / Express
+- **Active Rules:** 71 canonical security rules enabled in `.torusguard/rules/active/`
+- **Configuration:** Written to `.torusguard/config/torusguard.json`
+- **Security Policy:** Baseline `SECURITY.md` generated
+- **Next Step:** Run `npx torusguard audit` or `/torusguard audit` to scan for flaws
+```

package/skills/torusguard-recheck/SKILL.md

@@ -1,54 +1,77 @@
----
-name: torusguard-recheck
-description: Execute targeted differential AST re-scan against modified files, verify fix closure, and assert zero regressions.
-version: 1.0.0
-workflow: .torusguard/workflows/recheck.md
-tools: Read, Grep, Glob, Write
-scripts-binding:
-  - .torusguard/scripts/run_manager.py
-  - .torusguard/scripts/finding_scorer.py
----
-
-# TorusGuard Recheck — Targeted Differential Audit & Regression Verification
-
-## Objective
-Re-scan only files modified by remediation patches, confirming that targeted vulnerability sinks have been neutralized and verifying that no new security regressions were introduced.
-
----
-
-## Execution Steps
-
-1. **Identify Modified Files:** Read `apply-log.json` from active run directory.
-2. **Handle CLI Failures:** If `npx torusguard recheck` crashes or reports a false negative, YOU must act as the manual arbiter.
-3. **Execute Differential Scan:** Re-run active `TG-*` rules exclusively against modified files. If doing this manually, use `view_file` to read the patched file and cross-reference with the active rules.
-4. **Evaluate Finding Status:** Transition finding state according to transition rules below.
-5. **Assert Regression-Free:** Ensure no new security warnings triggered on altered lines.
-6. **Update State & Memory:** Record status in `findings.json` and sync verified results to memory engine.
-7. **Emit Recheck Report:** Save summary in `.torusguard/runs/<run_id>/recheck-report.md`.
-
----
-
-## State Transition Rules
-- **`Fixed`:** Targeted vulnerable AST pattern is absent and no new issues appear on modified lines.
-- **`Partially Fixed`:** Vulnerability surface was reduced but sanitization remains incomplete.
-- **`Not Fixed`:** Vulnerable AST sink remains present in patched file.
-- **`Regression`:** Patch introduced a new security rule violation on modified lines.
-
----
-
-## Safety Constraints
-- Restrict AST scan strictly to modified files and direct callers.
-- If regression is detected, halt and recommend immediate rollback via `.bak` snapshot. If CLI rollback fails, you MUST manually revert the file using your write tools.
-- Read-only differential analysis.
-
----
-
-## Output Format
-```markdown
-✅ [TorusGuard] Differential Recheck Completed (AI Assisted)
-- Modified Files Scanned: <Count>
-- Target Finding: <Finding ID> ──► FIXED
-- Regressions Detected: 0
-- Report: `.torusguard/runs/<run_id>/recheck-report.md`
-Next: Run `/torusguard report` to export final SARIF and release summary.
-```
+---
+name: torusguard-recheck
+description: Execute targeted differential AST re-scan against modified files, verify fix closure, and assert zero regressions via CLI or AI Agent.
+version: 1.3.4
+workflow: .torusguard/workflows/recheck.md
+tools: Read, Grep, Glob, Write, run_command
+scripts-binding:
+  - .torusguard/scripts/recheck_runner.py
+  - .torusguard/scripts/audit_runner.py
+  - .torusguard/scripts/memory_engine.py
+  - .torusguard/scripts/term_ui.py
+---
+
+# TorusGuard Recheck — Targeted Differential Audit & Fix Closure
+
+## Objective
+Execute differential security re-scans strictly scoped to modified files and adjacent boundaries, verifying that targeted vulnerability patterns have been eliminated and asserting zero security regressions.
+
+---
+
+## Two Execution Modes
+
+### Mode A: Automated CLI Execution
+Run the differential recheck engine from the terminal:
+```bash
+# Recheck the latest applied run
+npx torusguard recheck
+
+# Recheck a specific project directory
+npx torusguard recheck ./my-project
+
+# Recheck a specific run ID
+npx torusguard recheck --run run-20260910-121618-audit
+
+# Machine-readable JSON output
+npx torusguard recheck --json
+```
+**Under the Hood:** Executes `python .torusguard/scripts/recheck_runner.py`.
+- Evaluates applied candidate bundles from `.torusguard/runs/<run_id>/bundles/`.
+- Executes single-file targeted differential AST scans via `audit_runner.scan_file()`.
+- Calculates formal status transitions:
+  - `✔ [Confirmed Fixed]`: Vulnerable pattern absent, zero regressions.
+  - `✖ [Regressed]`: New security finding introduced by patch.
+  - `⚠ [Unresolved]`: Vulnerable sink still present.
+- Emits run-level transition artifact: `.torusguard/runs/<run_id>/recheck.md`.
+- Synchronizes verification telemetry to `.torusguard/memory/events.json`.
+- Displays 75-column terminal cards.
+
+### Mode B: In-Session AI Chat Agent Differential Scan
+When evaluating patches directly in AI chat:
+1. **Identify Modified Files:** Read `diff_summary.md` or git status for files modified in the active run.
+2. **Re-Scan Sinks:** View the target file and verify that the specific rule violation (e.g. `TG-INPUT-003`, `TG-PLATFORM-001`) is no longer triggered.
+3. **Assert Zero Regressions:** Verify that no new vulnerabilities (like raw concatenation or unvalidated input) were introduced by the fix.
+4. **Update Status:** Log outcome in `.torusguard/runs/<run_id>/recheck.md`.
+5. **Memory Telemetry:** Record verification event with `memory_engine.record_event("fix_verified", ...)`.
+
+---
+
+## Recheck Status Transitions
+| Outcome | Visual Indicator | Meaning | Required Action |
+| :--- | :--- | :--- | :--- |
+| **Confirmed Fixed** | `✔ [Confirmed Fixed]` | Sink eliminated, zero regressions | Proceed to Posture Report |
+| **Unresolved** | `⚠ [Unresolved]` | Flaw still present in file | Re-harden with alternative pattern |
+| **Regressed** | `✖ [Regressed]` | New security flaw introduced | Instant `npx torusguard rollback` |
+
+---
+
+## Output Card Format
+```markdown
+### ✅ TorusGuard Differential Recheck Completed
+- **Run ID:** `run-20260910-121618-audit`
+- **Confirmed Fixed:** 2 vulnerabilities verified closed
+- **Regressions:** 0 detected
+- **Unresolved:** 0
+- **Artifact:** `.torusguard/runs/<run_id>/recheck.md`
+- **Next Step:** Run `npx torusguard report --html` for visual dashboard
+```