n2-fieldops 1.1.0

Feature release: skill-based Prompt Decorator runtime with the +++N2 and +++Storm
meta-decorators, and a token-optimized Codex profile.

New

Meta-decorators +++N2 (deep rigorous analysis) and +++Storm (novel-idea
generation): each treats +++ChatScope as active first, then activates its defined
decorator stack. Full expansion, parameters, and deactivation rules ship in the
fieldops-prompt-decorators skill at references/meta-decorators.md.
The decorator catalog, state machine, composition rules, docs reference, and the
deterministic parser (scripts/parse_decorators.py) all validate and document
N2/Storm parameters (iterations, limit, diversity, lens, domain).

AGENTS.md (Codex profile) is now skill-based and token-optimized (417 -> 189
lines): the decorator reference moved into the fieldops-prompt-decorators skill,
while runtime awareness keeps natural-language activation, decorator precedence,
one-line acknowledgements, and the mandatory [Active: ...] state line with a
re-read rule after context compaction.
Restored profile sections: Scope and Cooperation, Reverse Engineering and Malware
(11-step block), Engineering Quality and Token Efficiency, and Idea Generation and
Novel Concepts.
claude-plugin-template/docs/CLAUDE_PROFILE.md gained the same decorator-runtime
capabilities (natural-language activation, meta-decorator expansion, the
[Active: ...] state line, N2/Storm parameter limits) and the same uncensored
stance as the Codex profile: full Scope and Cooperation language, sandbox-internal
network identifiers, and CTF targets treated as authorized sandbox assets by
default. The 1.0.x hardening of authorization and network assumptions is reverted.

The canonical repository remains one root-level tree with exactly 17 SKILL.md
directories. Install from CC Switch with:

Repository URL: https://github.com/download4you/n2-fieldops
Branch: main

The native Claude plugin and marketplace bundles are generated from the same
canonical tree.