tricritical
v1.0.0by nisavid · author: Ivan D Vasin
Tricritical: independently planned intent, runtime, and structure critics, adjudication, authority-gated revision, and one fixed-point loop.
What tricritical does, in the publisher's words
Tricritical is an Agent Plugins review package with native Claude agent aliases. It separates independent observation from adjudication, authorized revision, and repetition.
Root plugin.json and skills/ are the canonical Agent Plugins package. The package-level references/ files and topology.json are the authored shared sources. Each public skill loads generated byte-identical projections beneath its own references/ directory, so a client that installs one skill root can resolve every shared runtime resource without package-parent access. Run the Tricritical content-lock writer after changing an authored shared source; it refreshes the required per-skill projections before locking them. Claude Code has no standard agent component, so its native adapter auto-discovers root agents/*.md; each file is an exact minimal forwarder to one public skill. Codex consumes skills/ and each skill's agents/openai.yaml, but not root Claude agents. A Codex named-agent projection belongs to a local harness configuration and must call the public skill rather than copy its policy.
From the project README, punctuation lightly normalized · Open the README on GitHub
- Skills
- 7
- MCP servers
- 0
- Stars
- 0
- License
- MIT
- Repo created
- 2026-09-01
- Last pushed
- 2026-10-21
- Publisher type
- User
- Version
- 1.0.0
Links
Skills · 7
- adjudicate
- Use when Tricritical or external review findings need evidence-based acceptance, rejection, deferral, or follow-up classification without changing the reviewed candidate.
- intent
- Use when reviewing a candidate for requirement fit, missing outcomes, scope, authority, acceptance criteria, or alternate user and operator perspectives.
- loop
- Use when the user asks to review and revise a candidate until the current increment is done, or when an authority-gated fixed-point review cycle is explicitly required.
- review
- Use when a candidate needs an independent read-only review, a critic subset, or a synthesis of intent, runtime, and structure evidence without edits.
- revise
- Use when accepted Tricritical findings must be integrated into a candidate under explicit original mutation authority and a bounded review scope.
- runtime
- Use when reviewing a candidate for runtime correctness, security, compatibility, failure paths, concurrency, operational behavior, or false verification.
- structure
- Use when reviewing a candidate for needless complexity, duplication, weak ownership, boundary leaks, over-abstraction, or opportunities to delete and simplify.
Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.
Category
Agent Tooling. Assigned by this directory. Agent Plugins 1.0.0 has no category field, so no manifest declares one.
Keywords
review · intent · runtime · structure · adjudication · revision
Related by publisher and keywords
Proseweaving: house standards and verification for agent-authored human-facing prose.
Clear, adopt, maintain, replace, and retire exact third-party software components under explicit policy.
Mergecraft: Issue: PR relations, pull-request authoring, publication, Graphite transport, feedback, readiness, merge, and stack repair.
Rolecasting: portable model selection and delegation topology for agent work.
Versionkeeping: safe Git checkpoints, conflict resolution, exact-lease publication, provenance-aware worktree cleanup, and history-preserving fork sync.
HTML is the new markdown. Lavish is the new editor for your HTML artifacts.
[](https://agentpluginsdirectory.com/plugins/tricritical--nisavid)✓ Verified . We fetched the manifest from GitHub and checked it against the official Agent Plugins 1.0.0 schema at agent-plugins.org.