sdl-mcp-agent-workflow

Use when working in an SDL-MCP-enabled repository, including repository exploration, task context, code inspection, runtime execution, edits, or SDL-MCP tool calls.

Jump to install

Repository: GlitterKill/sdl-mcp
Last source activity: August 25, 2026 at 16:36
Detected SKILL.md language: English
Stars: 467
Forks: 29

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

A direct command skips the review prompt. Inspect the source before running it.

npx skills add https://github.com/GlitterKill/sdl-mcp --skill sdl-mcp-agent-workflow

The command stays on one line. Scroll horizontally to inspect it before copying.

Prefer a local copy? Download the files currently available to SkillsMP.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer

2 files

SKILL.md

Source instructions · Read-only preview

Back to search Back to installation

SDL-MCP Agent Workflow

Use SDL-MCP as the repository boundary.

When a client bridge returns the raw MCP result, prefer structuredContent; use text content only as a fallback for older servers. Do not emit both or return the whole MCP response envelope to the agent.

Start with repo.status.

Use sdl.context for task-shaped explain, debug, review, and implement work. Its request is flat and requires budget.maxTokens; never send options, contextMode, or answerFirst.

Use sdl.retrieve for one card, slice, skeleton, hot path, or justified code window. Never use file.read for indexed source.

Use sdl.workflow for runtime execution, transforms, dependent calls, and batch mutations. Persist command output and query only needed failure lines. runtimeExecute executes repository tooling. Permitted uses include build, test, lint, compiler, named scripts, and targeted edit scripts. Do not use it to inspect, search, or print repository files. Use sdl.context or sdl.retrieve for indexed source and sdl.file with op="read" for other files. Do not guess runtimeQueryOutput arguments; replay a returned action unchanged or call focused sdl.manual for runtime.queryOutput first.

Read non-indexed files through sdl.file. Its targeted write operation can update one indexed file with live reconciliation; prefer symbol or search-edit preview/apply operations when they can anchor the change.

Keep responseMode: "auto" for potentially large results. When a result returns a canonical response.get continuation (nextAction or action) for sdl.retrieve with op: "responseGet", replay its returned action and arguments unchanged; do not reconstruct it. Outer repoId owns trusted dispatch, and detail/includeDiagnostics stay outer sdl.retrieve controls. Nested args contains only artifact view and paging fields; nested repoId is invalid. Use workflow responseGet only when direct sdl.retrieve is unavailable or an existing multi-step workflow needs it.

Reuse refs and ETags. Set refsMode: "off" only for complete or byte-stable output.

Never call index.refresh, directly, through sdl.workflow, or via sdl-mcp index, without explicit user approval in the current turn. Dirty semantic flags, graph verification, parser-state/provenance warnings, and refresh recommendations are diagnostics, not approval. Do not wait for semantic freshness or verification unless the task requires latest-revision graph proof. Use an SDL file-based edit fallback after provenance failures.

Call usage statistics only when the user asks for token savings or telemetry.

Use tool recipes for exact v2 request shapes.