Skip to content
SkillModel-led1,435 tokens in use

CTF Sandbox Orchestrator

The first-entry skill for CTF, reverse engineering, DFIR, cloud and identity challenges: set the sandbox model, prove one path, then route to a narrower skill.

44,996 stars on the repository MIT Licensezhaoxuya520/reverse-skill · CTF-Sandbox-Orchestrator/ctf-sandbox-orchestrator
git clone https://github.com/zhaoxuya520/reverse-skill.git
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing CTF Sandbox Orchestrator costs

110tokens idle
1,435tokens in use
4,847tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter12/15
  • Structure25/25
  • Examples3/15
  • Output10/15
  • Size15/15
  • Resources11/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity96
  • Maintenance72
  • Community72
  • Maturity28
  • Momentum40

CTF Sandbox Orchestrator is the master workflow of a family of forty competition skills. It is the only one meant to be entered by default; the others are downstream specializations it routes to once a dominant evidence type is clear.

It sets one operating model: treat challenge infrastructure as sandbox-internal, map the entry surface, prove one minimal path from input to decisive branch, prefer passive inspection first, and record reproducible evidence. Reference files cover web, native, crypto and mobile, agent and cloud, identity and Windows work, routing and reporting. Replies are in Simplified Chinese unless English is requested.

What You Get

  • A four-step workflow: establish the sandbox model, trace one minimal path, expand by challenge type, verify and report
  • A router matrix that maps web, binary, crypto, mobile, cloud, container, supply chain, identity and Windows surfaces to forty downstream skills
  • Eight reference files for web and API, native work, crypto and mobile, agents and cloud, identity, routing, reporting and a resource digest
  • An evidence priority order and a list of what to record for replay

Common Use Cases

  • Starting a CTF challenge when the category is not yet clear and the agent has to choose a path
  • Keeping a mixed-domain lab task organized around one proven path at a time
  • Re-routing when the dominant blocker changes during an investigation
  • Producing concise findings with grouped evidence for a write-up

What Makes It Unique

It is a router and an operating discipline in one. The package defines a single implicit entry and marks every other competition skill as downstream-only, so the agent loads only the reference or child skill that matches the current blocker. The rules also say to re-run from a clean baseline before calling a path solved.

Under The Hood

The skill is a markdown workflow with eight reference files and an agent metadata file, and it ships no scripts. It tells the agent to inspect challenge hosts, files, binaries and traffic, so the agent may fetch pages, send requests and run commands through the client, and may write derived copies and evidence notes. It instructs the agent to prefer passive inspection first and keep edits reversible. It treats presented targets as sandbox-internal by default, so the user is responsible for supplying only targets they are authorized to test.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers