agentleFS
WingedGuardianWingedGuardian/skyrimvr-claude-toolkit/.agents/skills/toolkit-housecarl/SKILL.md

toolkit-housecarl

Skill22 starsChanged 4 days agoMITView on GitHub

Query Skyrim plugin records, compare winners and fields, inspect SkyPatcher/config references, or author patches and ordinary plugin edits through the installed House Carl gateway. Use for supported SE/AE/VR data-layer work; use DevBench for observed runtime behavior and existing toolkit tools for saves, mesh authoring and compilation.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run python tools/housecarl.py doctor first when context is unclear.

    line 9
  2. Very specific:

    Use the registered housecarl gateway, or the same operation through:

    line 14
  3. Very specific:

    Put the tool's arguments, not a JSON-RPC envelope, in the request file.

    line 20
  4. Distinctive:

    Do not copy those separately licensed corpora into toolkit source or a release.

    line 29
  5. Very specific:

    Use plugin-qualified XXXXXX:Plugin.esp identities.

    line 31
  6. Distinctive:

    Keep the distinction between file state, modeled SkyPatcher effects and observed in-game state.

    line 32

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- GENERATED by scripts/gen_agent_trees.py from agent/skills/toolkit-housecarl/SKILL.md. Toolkit contributors: edit the source. Your own notes belong in KNOWLEDGEBASE.local.md, which updates never replace. -->

House Carl through the toolkit

Run python tools/housecarl.py doctor first when context is unclear. It names the configured authority, exact source and capabilities. Wrye-managed order comes from its persisted plugins.txt, with the game's implicit-master and master-block rules. Deployment ownership is a separate fact. Never repair or replace the user's order to make the backend agree.

Use the registered housecarl gateway, or the same operation through:

python tools/housecarl.py call housecarl_records --arguments-file <request.json>

Put the tool's arguments, not a JSON-RPC envelope, in the request file. CLI output carries context and result; the latter contains the upstream MCP content. A refusal exits 2. Consult tool schemas before constructing calls. project.form=tree compares overrides; chain is for graph walks.

For a record schema or Papyrus/distributor grammar, locate the installed package from doctor or the gateway configuration at .skyguard/housecarl/config.json: the server's parent directory is beside skills/. Read the relevant upstream SKILL.md and search only the required references. The seven specialist directories are mutagen-reference, papyrus-reference, skypatcher-authoring, spid-authoring, kid-authoring, open-animation-replacer and skse-plugin-authoring. Do not copy those separately licensed corpora into toolkit source or a release.

Use plugin-qualified XXXXXX:Plugin.esp identities. Runtime addresses are unqualified here. Keep the distinction between file state, modeled SkyPatcher effects and observed in-game state. An unresolved reference in a commented example is not an active broken setting. Facegen provenance and native script/DLL pairing are unavailable until their provider/runtime information is qualified.

For authorized ordinary edits, use apply, create, remove, forward, copy or create_plugin. Specify a new patch by default. Existing-plugin in_place names the exact target. The gateway backs up, stages and independently validates edits before publication; upstream acknowledgement does not replace that protection. Do not invoke the raw server to bypass a refusal. Existing-file edits change the active file only, so manager redeployment may replace them. New patches are staged and require manual activation. Never edit activation files on the user's behalf.

Inspect the returned receipt and changed records. Keep its backup permanently. Restore with python tools/housecarl.py restore <transaction-id>; a changed destination refuses rather than overwriting later work. Compaction, merges and other identity-renumbering operations are outside this integration. Existing tools retain compilation, archive writes and NIF authoring.

Installation, supported providers and qualification limits: docs/housecarl.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.