Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Provider-neutral planner for synchronizing MCP configuration, skills and rule files across agent providers and app classes. It discovers evidenced local options and lets the user choose truth, targets, direction and conflicts.
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -69% | 0% |
<img src="banner.png" width="100%" alt="agent-config-sync banner">
> Deutsch — Offizielle Deutsch-Version / Documento Oficial en Deutsch.
The skill separates endpoint selection, resources and truth. Run:
bashpython scripts/sync.py --discover python scripts/sync.py --offer
The user can select an explicit endpoint list, one provider across app classes, one app class across providers, or all detected endpoints. Detection is evidence, not authorization.
Truth can be one endpoint, one file, an ordered set of files such as multiple AGENTS.md layers, or a skills directory. No filename or provider is the implicit hub. Without a selected truth source, plans stay blocked.
Review --status and --plan; only use --apply --yes after approval. MCP blocks and skill directories are implemented. Rule-file topologies remain fail-closed until the user selects a merge/redirect adapter.
Other measured skills in the registry, with their headline benchmark lift.