mantis-launch

A launcher for automated security reviews of source files or whole code repositories. It starts a process that looks for vulnerabilities, checks whether they can be exploited, reproduces crashes, creates possible patches, and assesses risk.

In plain words
What is it for?
Use it to review a file or repository for vulnerabilities, validate suspected issues, test exploit or crash behavior, generate patches, and calibrate risk.
Why use it?
It brings several security-review steps into one campaign instead of requiring separate manual runs. Preflight checks and environment setup help the review start with the required configuration.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/google/mantis/mantis-launch
Any agent
npx skills add google/mantis --skill mantis-launch
Clone the repo
git clone --depth 1 https://github.com/google/mantis

Made for: Claude Code, Codex.

Per session 60 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,156 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00060 $0.01156
Opus 5 $0.00030 $0.00578
Sonnet 5 $0.00012 $0.00231
Haiku 4.5 $0.00006 $0.00116

Measured 2d ago against content hash 92eff8c3f8ba, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mantis-launch scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

reference/skills/mantis-launch/SKILL.md · 127 lines

How it starts

The opening of the file, as written. The whole thing — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Campaign Launcher (/mantis-launch)

System Goal

Autonomous Security Review Pipeline Launcher. Initiates end-to-end vulnerability discovery, independent verification, exploit viability analysis, crash reproduction, automated patch generation, and risk calibration campaigns across a target file or entire codebase repository.

Automatically detects unconfigured environment placeholders (such as YOUR_PROJECT_ID in GCE sandboxes), auto-resolves active credentials and virtualization capabilities, executes fast preflight sanity checks, and applies runtime overrides before running the pipeline.

Command Definition

  • Command: /mantis-launch
  • Description: Launches automated multi-agent vulnerability discovery and validation campaigns.
  • Execution Commands:
    # From reference/ directory:
    ./run.sh <target_file_or_dir> [flags...]
    # or: python3 scripts/launch.py <target_file_or_dir> [flags...]
    
    # From repository root:
    ./reference/run.sh <target_file_or_dir> [flags...]
    # or: python3 reference/scripts/launch.py <target_file_or_dir> [flags...]
    
  • CLI Options:
    • target (positional): Path to a single source file (e.g. src/auth.py) or a root repository directory (e.g. . or /path/to/repo).
    • --sandbox / -s: Override sandbox mechanism (static-only, gvisor, microsandbox, gce).
    • --model / -m: Override AI model (e.g. gemini-3.7-flash, vertex_ai/claude-opus-5, vertex_ai/zai_org/glm-5.2-maas, openai/{MODEL_ID}).
    • --api-base: Custom endpoint URL for OpenAI-compatible LLM deployments (e.g. http://localhost:8000/v1).
    • --reasoning-effort: Reasoning effort level (low, medium, high).
    • --timeout: LLM request timeout in seconds.
    • --db / -d: Custom path to SQLite knowledge database (default: knowledge.db).
    • --workflow / -w: Path to custom workflow.json layout definition.
    • --preflight-only / --test / --preflight: Run preflight checks and exit without starting the campaign.
    • --interactive: Launch interactive configuration wizard before execution.
    • --dry-run: Display launch plan and indexed files without calling AI models.
    • --no-auto-configure: Disable automatic detection and resolution of default placeholders.

Read the full file on GitHub · 127 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 127 lines · 60 tokens per session scan A 92eff8c3f8ba

Subscribe to this mod's changes

mantis-launch is a skill published in the GitHub repository google/mantis (781 stars, last pushed 4d ago), licensed Apache-2.0. It adds 60 tokens to every session and 1,156 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.