メニュー
Create, write, list, and pull versioned exploit artifact folders through the kartifact CLI.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Wrap an ordinary C exploit or PoC in kernel_PoCs/common so one source project can build as a Linux or Android static executable, dynamic executable, constructor-loaded shared object, or JNI library with local, listening, or reverse-shell root payloads. Use when creating a common-compatible exploit project, converting a C program with main(), setting up exploit entry and post-escalation payload code, repairing its Makefile or entrypoint integration, or initializing the bundled example project.
Add a Linux or Android kernel exploit primitive to ingots_tools/primitives/src/primitive_data in the exact format consumed by primitives/src/primitives/__init__.py. Use when creating a new primitive entry with metadata.json, PRIMITIVE.md, and primitive.c, or when updating an agent to generate primitives for this repository.
Use the android-kdebug skill when debugging Android apps over ADB. The android-kdebug skill documents using kdebug Frida CLI, the kdebug LLDB CLI, or common adb commands.
Use this skill to launch and manage Android Cuttlefish instances using the cuttle-cli and cuttle-server.
Use the high-level object_db wrapper API to scan kernel objects in Python by loading an ObjectSet and filtering HeapObject entries with helper predicates instead of raw SQL.
Use the `testvm` CLI to pack and unpack initrds, build the default BusyBox initrd, and boot Linux kernels under QEMU on `x86_64`, `arm`, or `aarch64`. Use this when the user asks to run, debug, or explain `testvm` commands.
| name | kartifact |
| description | Create, write, list, and pull versioned exploit artifact folders through the kartifact CLI. |
Use kartifact when an unmanaged agent needs to exchange an artifact folder with the Ingots artifact database. The artifact type must come from the task or that type's documentation; kartifact intentionally has no type-discovery command.
uv run kartifact.artifactdb.sqlite or anything under the managed blobs/
directory directly.artifact.id in artifact.toml as the checked-out revision token.artifact.parent_id; it is rendered provenance.create and pull.--json when consuming output programmatically. It is a global option
and must precede the command.uv run kartifact create <type> <folder> --name <safe-name>
The name may contain letters, digits, ., _, and -. Edit the generated
artifact.toml metadata and add the artifact's files before writing it.
uv run kartifact --json write <folder>
A successful write creates an immutable database revision and rewrites the
working folder's artifact.toml with its new id and parent_id. Keep using
that rewritten folder for later revisions.
If a write reports a stale-head conflict, another revision already advanced the
same type/name. Either pull the newer revision and reapply the work, or choose a
new unused name in artifact.toml to create an intentional fork. Never remove
or replace the ID to bypass a conflict.
If kartifact reports source_update_failed, the returned artifact ID was
committed even though the working TOML could not be rewritten. Pull that ID into
a new empty folder before continuing.
uv run kartifact list <type>
uv run kartifact --json list <type>
uv run kartifact --json list <type> --include-shadowed
The default list contains visible name heads. Use --include-shadowed when an
older immutable revision is needed.
uv run kartifact --json pull <revision-id> <empty-folder>
Pull always selects by UUID, copies that revision's files, and renders canonical
database metadata into artifact.toml.