kimi-code-plugin-cc

Plugin kimi-code-plugin-cc from marketplace kimi-code-cc in ArchiDoxx/Kimi-code-Plugin-CC. Commands read from the repo's manifest — 1 month ago.

Skills (10)

• bridge

  One-shot call to a headless CLI agent (Kimi Code by default) via the bridge MCP server. Use for a single question or task that does NOT need an iterative loop. Lowest-overhead way to get an external agent's answer.

• code-review

  Focused single-pass external code review. Sends code + a structured review brief to Kimi Code and returns concrete, prioritised findings. The daily-workhorse skill for reviewing a file, function, or diff. Use /kimi-code-review <target>.

• contract-audit

  External audit of a frozen API / data-model contract against its implementation. The architect owns the team — this skill sends the frozen contract AND the implementing code to Kimi Code and gets back field-by-field drift, missing pieces, and fail-safe coverage gaps. Single-pass, structured. For arc

• fail-safe-audit

  External audit of every failure / stale / fault path for safe-state behavior (never the safe / "all-clear" state on failure). Sends the code to Kimi Code and gets back each path that can reach an unsafe state. Single-pass, structured, empty-output-safe (never reports OK on failure). For devs and arc

• planning-loop

  Iteratively build or refine a plan with an external agent, refining up to a max-iteration budget. Use when a complex task needs a structured plan from a second perspective before implementation.

• review-loop

  Multi-round external code review that iterates up to a max-iteration budget, stopping early on 'approve'. Use when one review pass is not enough and you want the reviewer to refine its findings against the same target.

• santa-loop

  Adversarial dual-review. Fail-closed — returns green ONLY when two independent reviewers both approve. Use for security-critical or safety-critical code where a false 'looks fine' is costly.

• second-opinion

  Quick external second opinion on a design decision, approach, or trade-off. One-shot call to Kimi Code with a decision-focused prompt. Use /kimi-opinion "<question>" for fast design/architecture feedback before you commit to an approach.

• team-design-review

  External review of a module boundary or interface design — coupling direction, dependency rules, deep-module quality, abstraction leaks. Send the proposed design to Kimi Code and get back concrete, prioritised design findings. Single-pass, structured. For architects shaping module boundaries.

• test-gap-audit

  External audit of TESTS for a target — finds missing cases (edge cases, error paths, fail-safe paths, documented incidents) rather than reviewing code. Sends the code UNDER test plus its tests to Kimi Code and gets back concrete coverage gaps. Single-pass, structured. For devs preparing a PR against

Slash commands (4)

• /kimi-code-plugin-cc:kimi-code-review

  Focused single-pass external code review of a file or inline code by Kimi Code. Daily-workhorse review command — concrete, prioritised findings in one shot.

• /kimi-code-plugin-cc:kimi-opinion

  Quick external second opinion from Kimi Code on a design decision, approach, or trade-off. Decisive, trade-off-aware answer in one shot.

• /kimi-code-plugin-cc:kimi-review

  Review a target with a single-agent loop, or an adversarial dual-review (santa) where Claude is the heterogeneous second reviewer. For a single fast pass use /kimi-code-review instead.

• /kimi-code-plugin-cc:kimi-run

  Run a single prompt through a registered headless CLI agent (Kimi by default) via the bridge MCP server. Lowest-overhead one-shot call.

Subagents (2)

• bridge-orchestrator
• review-adversary

MCP servers (1)

• kimi-code-plugin-cc

  runs: uv run --project ${CLAUDE_PLUGIN_ROOT} kimi-code-plugin-mcp

Similar plugins