# PandaWard (mcpb · pandaward-mcp.mcpb)

Local, evidence-based application security assessment with an evidence ledger and retests.

- Trust score: 52/100 (low)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- mcpb · `pandaward-mcp.mcpb`: 52/100 (this document), [markdown](https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp.md), [page](https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp)

## Channel facts

- Registry: `mcpb`
- Package: `https://github.com/ratomirjovanovic/pandaward/releases/download/v0.2.13/pandaward-mcp.mcpb`
- Version: `0.2.13`
- Transport: `stdio`
- SHA-256: `39cd7b6d84aedd906741d7f0c0237e8698473bbdab8aacd306cddb944e0a5c8a`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Supply Chain Security**: 0/100
  - No malware scan is available for this kind of package: the supply-chain vendors we use do not cover it. This is a permanent gap in our coverage, not a finding about the package.
  - Known CVEs could not be checked: we found no dependency list to check in this bundle: either it runs a compiled binary, whose bundled manifests we do not read, or none of the package.json, requirements.txt, pyproject.toml or uv.lock files it ships for its runtime declares a dependency we can read (a requirements.txt that only points at a setup.py project, a local path, or a nested requirements file we cannot follow counts as none).
  - Install-script risk not yet assessed.
  - Dependency health could not be checked: we found no dependency list to check in this bundle: either it runs a compiled binary, whose bundled manifests we do not read, or none of the package.json, requirements.txt, pyproject.toml or uv.lock files it ships for its runtime declares a dependency we can read (a requirements.txt that only points at a setup.py project, a local path, or a nested requirements file we cannot follow counts as none).
- **Provenance & Transparency**: 94/100
  - Source repository is publicly reachable at the declared URL.
  - Cryptographically verified build provenance (signed, bound to ratomirjovanovic/pandaward).
  - Clear OSI-approved license (AGPL-3.0-only).
  - Maintenance history not yet determined.
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 66/100
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 1305 tokens (~87/item across 15 items; 15 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 27/100
  - Stability observed for 8 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 88/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 63% of tool parameters carry a description.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 15 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 16 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

**Unverified: 1 category.** A category scored 0 because we could not verify it: a data source with nothing on this package, evidence we could not reach, or a check we could not run. We only credit what we can confirm.

## Install

### How do I install the PandaWard MCP server?

PandaWard ships as an MCPB bundle, a single file you download and open in an MCP host that supports MCPB bundles, such as Claude Desktop. The host reads the launch command from the bundle's own manifest, so there is no command to copy. Before opening it, check the file's SHA-256 matches the one the MCP registry declares for it, 39cd7b6d84aedd906741d7f0c0237e8698473bbdab8aacd306cddb944e0a5c8a; the commands to do that are on this page. If it does not match, do not open it.

- Download bundle: `https://github.com/ratomirjovanovic/pandaward/releases/download/v0.2.13/pandaward-mcp.mcpb`
- SHA-256: `39cd7b6d84aedd906741d7f0c0237e8698473bbdab8aacd306cddb944e0a5c8a`

### Verify on macOS / Linux

```bash
echo "39cd7b6d84aedd906741d7f0c0237e8698473bbdab8aacd306cddb944e0a5c8a  pandaward-mcp.mcpb" | shasum -a 256 -c
```

### Verify on Windows (PowerShell)

```powershell
(Get-FileHash -Algorithm SHA256 .\pandaward-mcp.mcpb).Hash -eq "39cd7b6d84aedd906741d7f0c0237e8698473bbdab8aacd306cddb944e0a5c8a"
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-05 (score 52, 0)

- [functional regression] Schema quality: 78 → 87
- [functional regression] Maintenance: pass → unverified
- [functional] Package version: 0.2.12 → 0.2.13

### 2026-11-04 (score 52, 0)

- [security regression] Stability: 0.20 → unverified
- [security regression] Tool safety: pass → unverified
- [functional regression] Capabilities: pass → unverified
- [functional regression] Tool coverage: 100 → unverified
- [functional] First check of Schema quality: unverified
- [functional] Package version: 0.2.9 → 0.2.12
- [functional] Package version: 0.2.9 → 0.2.11
- [functional] Package version: 0.2.9 → 0.2.10

### 2026-11-03 (score 52, +1)

No change was recorded against any check on this day. Stability & Change Management went from 17 to 20. That category is still filling its 30-day observation window: 5 days of observed history at the previous scan, 6 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-01 (score 51, +1)

No change was recorded against any check on this day. Stability & Change Management went from 10 to 13. That category is still filling its 30-day observation window: 3 days of observed history at the previous scan, 4 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-31 (score 50, 0)

- [functional] Package version: 0.2.8 → 0.2.9

### 2026-10-30 (score 50, +1)

- [functional regression] Schema quality: 970 → 1074
- [functional] Package version: 0.2.7 → 0.2.8

### 2026-10-29 (score 49, 0)

- [security regression] Stability: 0.00 → unverified
- [security regression] Tool safety: pass → unverified
- [functional regression] Capabilities: pass → unverified
- [functional regression] Tool coverage: 100 → unverified
- [functional] First check of Schema quality: unverified
- [functional] Package version: 0.2.1 → 0.2.7
- [functional] Package version: 0.2.1 → 0.2.6
- [functional] Package version: 0.2.1 → 0.2.5
- [functional] Package version: 0.2.1 → 0.2.3
- [functional] Package version: 0.2.1 → 0.2.2

### 2026-10-28 (score 49)

First indexed and scored.

## MCP tools (15)

### `pandaward_get_status` (~60 tokens)

Compact assessment state from the local ledger: latest assessment, coverage split by how controls were decided, open findings, stale files, blockers, next action and the path to 100% (undecided controls grouped by the step that would decide them).

### `pandaward_list_coverage_gaps` (~57 tokens)

Applicable controls that were not decisively tested (BLOCKED, UNSUPPORTED, SUSPECTED, INCONCLUSIVE, ERROR), optionally for one domain A-R.

Input parameters:

- `domain` (string): Domain letter A-R

### `pandaward_get_control` (~37 tokens)

One control with its status, reason, references and evidence IDs.

Input parameters:

- `control_id` (string, required): e.g. C-001

### `pandaward_get_findings` (~59 tokens)

Findings from the ledger. Defaults to open findings. Snippets are redacted.

Input parameters:

- `confidence` (string)
- `control_id` (string): filter by control
- `limit` (integer)
- `status` (string)

### `pandaward_get_evidence` (~39 tokens)

One redacted evidence record after SHA-256 integrity verification.

Input parameters:

- `evidence_id` (string, required): ev_ followed by 16 hex characters

### `pandaward_get_graph` (~46 tokens)

Cross-domain edges (AB1, CG1, ...) with status, provenance and linked tests; optionally one edge.

Input parameters:

- `edge_id` (string): e.g. CG1

### `pandaward_list_entrypoints` (~79 tokens)

Entrypoints discovered by the latest assessment (HTTP routes with router prefixes resolved, GraphQL operations, gRPC methods, queue consumers, cron jobs), each with file, line and service.

Input parameters:

- `kind` (string)
- `limit` (integer)
- `service` (string): service root directory, e.g. services/api

### `pandaward_scan_passive` (~36 tokens)

Run the local passive scan (read-only inventory and static checks; no target network traffic). Writes only to the local ledger.

### `pandaward_retest` (~83 tokens)

Re-assess after code changes. Lists files changed since the last assessment, then re-runs only the adapters whose inputs changed (others are CARRIED with their previous decisions) unless changed_only is false. Local only; no target network traffic.

Input parameters:

- `changed_only` (boolean): re-run only affected adapters (default true); false runs the full passive scan

### `pandaward_build_report` (~29 tokens)

Render reports from the ledger into .pandaward/reports/.

Input parameters:

- `formats` (array)

### `pandaward_record_hypothesis` (~154 tokens)

Record a reviewer-proposed finding as HYPOTHESIS. PandaWard reads and redacts the cited lines itself; it is never CONFIRMED by this call.

Input parameters:

- `control_id` (string, required): catalog control the hypothesis belongs to
- `end_line` (integer)
- `file` (string, required): path relative to the project root
- `fix` (string): proposed remediation; max 4000 chars
- `proposer` (string): agent or reviewer name
- `rationale` (string, required): trust boundary, missing control, preconditions; max 4000 chars
- `severity` (string, required)
- `start_line` (integer, required)
- `title` (string, required): 5-160 characters

### `pandaward_propose_attestation` (~201 tokens)

Propose a PASS or FAIL verdict for a manual control (A-004, C-003, F-001, O-001, Q-001) after reviewing it. Evidence is required: file and line citations inside the project, each with why it shows the verdict; PandaWard reads, hashes and redacts the cited lines itself. The proposal is stored as pending and decides nothing: only a person accepts or rejects it with `pandaward attest <control> --accept|--reject --by <name>`, outside MCP. A FAIL verdict creates no finding; record the concrete problem with pandaward_record_hypothesis.

Input parameters:

- `command` (string): optional test or command that shows the verdict; max 500 chars
- `control_id` (string, required): manual control, e.g. C-003
- `evidence` (array, required)
- `proposer` (string): agent or reviewer name
- `verdict` (string, required)

### `pandaward_prepare_feedback` (~226 tokens)

Prepare a false positive or false negative report for the PandaWard project. Returns the report and a prefilled GitHub issue link for the human to review and submit; nothing is sent. By default the report holds only rule metadata and the file type; include_path and include_snippet add redacted location and code and need the user's consent.

Input parameters:

- `file` (string): false-negative: file with the missed issue, relative to the project root
- `finding_id` (string): false-positive: the PandaWard finding ID, e.g. PW-0007
- `include_path` (boolean): add the redacted file path and line (needs the user's consent)
- `include_snippet` (boolean): add up to 12 redacted lines of code (needs the user's consent)
- `kind` (string, required)
- `line` (integer)
- `note` (string): what happened, in the user's words; max 2000 chars
- `rule` (string): false-negative: rule or control that should have fired, e.g. D-003

### `pandaward_pending_feedback` (~96 tokens)

List the feedback reports this project has not sent yet, one per rule: rule findings a person closed as false positives, and reviewer findings a person confirmed where no rule reported the same control in the same file (missed issues). Also returns the user's feedback mode (ask, batch, off, or empty when not chosen yet). Nothing is sent: the user sends them with `pandaward feedback submit`, which asks for confirmation.

### `pandaward_request_active_test` (~29 tokens)

Explains how network-active testing is authorized. It never sends traffic and cannot approve scope.

## Diagnostics

Captured diagnostic sections: Provenance. The full working is on the page: https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp#diagnostics

## Score history

- 2026-11-05: 52
- 2026-11-04: 52
- 2026-11-03: 52
- 2026-11-02: 51
- 2026-11-01: 51
- 2026-10-31: 50
- 2026-10-30: 50
- 2026-10-29: 49
- 2026-10-28: 49

## Common questions

### What is the PandaWard MCP server?

PandaWard is an MCP server listed in the public MCP registry as io.github.ratomirjovanovic/pandaward. Local, evidence-based application security assessment with an evidence ledger and retests. This page covers its MCPB bundle (https://github.com/ratomirjovanovic/pandaward/releases/download/v0.2.13/pandaward-mcp.mcpb).

### Is the PandaWard MCP server safe to use?

PandaWard scores 52 out of 100 on VerifyMCP. Its build provenance is signed and verified. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the PandaWard MCP server expose?

PandaWard exposes 15 tools: pandaward_get_status, pandaward_list_coverage_gaps, pandaward_get_control, pandaward_get_findings, pandaward_get_evidence, and 10 more. Their descriptions and schemas cost roughly 1,231 tokens of context every time the server is loaded.

### What licence is the PandaWard MCP server under?

PandaWard declares the AGPL-3.0-only licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- Repository: https://github.com/ratomirjovanovic/pandaward
- Website: https://pandaward.dev/
- Changelog RSS feed: https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp.json
- HTML version of this page: https://verifymcp.io/servers/ratomirjovanovic-pandaward/pandaward-mcp