claude-code/skills/prompt-pack/SKILL.md
Generates a self-contained prompt pack from a spec. Embeds real code patterns from .vibeflow/ so any coding agent (Claude Code, Cursor, Copilot) follows the project's conventions without needing repo context. Use when handing off implementation to a separate session or agent.
npx skillsauth add pe-menezes/vibeflow prompt-packInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
What it does: Reads a spec from .vibeflow/specs/ (or path), builds a single prompt that includes objective, DoD, anti-scope, and real code patterns. Saves to .vibeflow/prompt-packs/<slug>.md. Give that file to the coding agent so it implements without needing the rest of the repo context.
Examples:
/vibeflow:prompt-pack .vibeflow/specs/login-flow.md — Generate prompt pack from the login-flow spec./vibeflow:prompt-pack login-flow — Same; agent finds the spec by name.The prompt pack MUST be written in the user's detected language. The opening line should be language-adaptive: Write the opening line in the user's detected language. The concept is: "You are only seeing this prompt; there is no context outside it."
All textual content (objective, DoD, anti-scope, guidance) should be in the detected language. Code, paths and technical names remain in English.
Generate a self-contained prompt pack for: $ARGUMENTS
Validate spec size. Read the spec and check:
.vibeflow/index.md or default ≤ 6 files)If EITHER condition is true, stop. Do NOT generate the prompt pack. Instead, inform the user: "This spec exceeds limits (N DoD checks / N files). Run /vibeflow:gen-spec again to split it into smaller specs before generating a prompt pack."
If $ARGUMENTS is a file path, read that spec file.
If it's a feature description, look for a matching spec in .vibeflow/specs/.
If no spec exists, generate one first (following gen-spec format), save it, then continue.
Read .vibeflow/ docs:
.vibeflow/conventions.md (always)## Pattern Registry YAML block from index.md (between
<!-- vibeflow:patterns:start/end --> markers). Cross-reference the
registry's tags and modules against the spec's scope. Load the top 3-5
matching patterns. If no registry exists, infer which ones are relevant.Read the actual codebase files relevant to this task.
Generate the prompt pack.
You are only seeing this prompt; there is no context outside it.
(Write this opening line in the user's detected language.)
Copied from the spec. Non-negotiable.
What NOT to do. Copied from spec.
Max files to change. Read the budget from the spec. If the spec doesn't
specify a budget, check .vibeflow/index.md for the Suggested budget
line. If neither is available, default to ≤ 6.
THIS IS CRITICAL. Embed the actual patterns the agent must follow.
Copy the relevant sections from .vibeflow/patterns/*.md and
.vibeflow/conventions.md directly into the prompt pack.
Include REAL code examples from the pattern docs — not references to external files the agent won't see. The prompt pack is self-contained.
Format as:
## Patterns to Follow
### <Pattern Name>
<description>
<real code example from this project showing the pattern>
### Coding Conventions
<relevant conventions for this task>
Real file paths from the codebase (verify they exist). Relevant code snippets for context.
Architectural direction, constraints to respect. NOT step-by-step prescriptive instructions.
Commands to validate the implementation. This section is REQUIRED.
.vibeflow/index.md or project config files (package.json,
pyproject.toml, Cargo.toml, etc.) to detect the test runner.npm test, pytest,
cargo test, go test ./...).## How to validate
1. Run tests: `npm test`
2. Verify manually: [description]
Which docs need changes after implementation.
If you cannot verify a file path exists, flag it with:
<!-- TODO: verify this path -->
Save the prompt pack to: .vibeflow/prompt-packs/<feature-slug>.md
Create the .vibeflow/prompt-packs/ directory if it doesn't exist.
After saving, suggest: "Prompt pack saved. Hand it off to the coding agent.
After implementation, run /vibeflow:audit with the spec to verify."
If this command is modified, update MANUAL.md to reflect the changes.
testing
Updates .vibeflow/ with corrections, new conventions, architectural decisions, or new patterns from natural language feedback. Also imports from external repos via --from <url|path>. Use to keep .vibeflow/ accurate as the project evolves.
development
Compiles statistics from audit reports: PASS/PARTIAL/FAIL rates, most violated patterns, most failing DoD checks, and quality trends. Use after running several audits to spot improvement areas.
data-ai
Fast-tracks small tasks into a prompt pack in one command. Skips discover, generates an ephemeral spec in memory. Use for well-defined tasks that fit in ≤4 files.
development
Generates a self-contained prompt pack from a spec. Embeds real code patterns from .vibeflow/ so any coding agent follows the project's conventions. Use when handing off implementation to a separate session or agent.