briefing
Briefing
Give the user a short, decision-ready view of the current task. Restore their task-scoped working context without dumping the whole conversation or codebase back at them.
The user is the decision-maker. Your job is to surface the relevant state, code, and constraints clearly enough that they can steer the next step.
Core Job
Restore the user's working context for the current task so they can make the next decision.
This skill is not for broad repository explanation. It is for reconnecting the user to the few details that matter now.
Core Behavior
- Work from the current request first.
- Treat the current task or question as the scope boundary.
- Read the current conversation before reading code.
- Pull in only the files and docs needed to explain the current state.
- Stay task-scoped.
- Focus on the current request or session, not the full project history.
- Include broader roadmap or open questions only when they materially affect the current task.
- Avoid generic repository summaries.
- Keep the briefing brief.
- Default to 8-14 bullets total, or the equivalent in short sections.
- Prefer high-signal facts over exhaustive detail.
- Do not walk line by line through code unless the user explicitly asks.
- Be mostly descriptive.
- Explain the current state, relevant code, and constraints.
- Light recommendations are allowed when they help orientation, such as the likely next step or the main thing to watch.
- Do not turn the briefing into a plan review, code review, or design critique unless the user asks for that mode.
Sources
Read sources in this order and stop as soon as the current task can be explained confidently:
- the current conversation
- the smallest set of relevant files in the codebase
- relevant local docs such as
AGENTS.md,ARCHITECTURE.md, or nearby module docs when they change how the task should be understood
Use file paths when they materially help the user reconnect to the code.
Briefing Styles
Default to a status briefing.
Do not depend on markdown headings or a rigid internal mode switch to decide behavior. Instead, infer the user's emphasis from their explicit request:
- If they ask to catch up, brief current state.
- If they ask what you are likely to do, emphasize likely implementation approach.
- If they ask to teach the codebase, emphasize roles, flows, and file relationships.
- If they ask for help deciding, emphasize constraints, tradeoffs, and the facts that matter now.
Treat these as variations of one job: produce a concise task-scoped briefing. If the request is mixed, combine the needed emphases without turning the response into a long menu.
Non-Goals
Do not use this skill to produce:
- a full architecture summary for the repository
- a code review or design review
- a project roadmap reconstruction unless the current task depends on it
- a verbose recap of the whole conversation
- a line-by-line walkthrough of implementation details unless the user explicitly asks
Workflow
- Identify the user's requested emphasis.
- If the user explicitly asks for current state, use a
statusemphasis. - If they ask what you will likely do next, use a
planemphasis. - If they ask to understand the relevant code, use a
teachemphasis. - If they ask what matters for a decision, use a
decisionemphasis. - If the request is ambiguous, default to
status.
- Recover the task boundary.
- State the current request in concrete terms.
- Distinguish the present task from earlier side discussions if needed.
- Gather only the relevant context.
- Start with the current conversation.
- Read the smallest set of files and docs that explain the current state.
- Prefer entry points, touched files, nearby interfaces, and stable docs over broad exploration.
- Synthesize for fast orientation.
- Collapse raw findings into the few facts the user needs now.
- Name the relevant files and modules directly.
- Explain why each one matters to the current task.
- Trim aggressively.
- Remove repeated points.
- Remove code trivia and low-signal history.
- Remove speculative statements unless they are clearly labeled.
Output
Default to short sections with flat bullets.
Use this shape unless the user asks otherwise:
Current Task
- One to three bullets on what is being worked on.
Relevant Code
- The key files or modules, with a short note on why each matters.
What Matters Now
- The most important implementation facts, behavior, or constraints for this task.
Decision Context
- Only the facts the user needs to make the next decision.
- Include a light next-step note only when it helps.
Adjust emphasis based on the user's request:
status: center the current implementation state.plan: center the likely approach and touched areas.teach: center roles, boundaries, and flow.decision: center constraints, tradeoffs, and high-impact unknowns.
Communication Rules
- Prefer the smallest useful briefing over a comprehensive one.
- Prefer concrete file references over abstract architecture language.
- Prefer "what matters now" over "everything that exists."
- Keep explanations compact but complete enough for decision-making.
- Do not repeat the conversation back to the user.
- Do not pad the output with generic advice.
- Do not create open questions unless the current task genuinely centers on them.
- If evidence is partial, say what is known and what is inferred.
Example Triggers
- "Use $briefing and catch me up on what matters for this task."
- "Before you implement this, give me a briefing on the relevant code."
- "Brief me on the current state of this feature."
- "Teach me the part of the codebase related to this request."
- "Give me a decision briefing for this change."
More from sjunepark/custom-skills
summarize
Use the steipete/summarize CLI to summarize URLs, local files, stdin, YouTube links, podcasts, and media with LLM models. Trigger when users ask to install or run summarize, configure model/provider API keys, tune output flags (length/language/json/extract/slides), set defaults in ~/.summarize/config.json, or troubleshoot summarize CLI errors.
41skills-cli
Operate the skills CLI to discover, install, list, update, remove, and initialize skills for Codex, Claude Code, and Pi. Use when users ask to manage skills from skills.sh, restore from lock files, sync skills from node_modules, or troubleshoot agent/installation scope (project vs global).
37post-implementation-review
Manually review already-implemented code for design flaws, abstraction issues, structural problems, or refactors that only became clear in real code. Use only when the user explicitly asks for a post-implementation review, explicitly asks whether recent implementation work revealed design or structure problems, or explicitly wants refactor recommendations after the code exists. Do not auto-trigger for ordinary implementation, debugging, explanation, or generic code review requests. Prefer embedded snippets with file-path comments over editor-oriented file and line references. Treat findings as signals about code shape and quality; prioritize root-cause design, ownership, abstraction, and organization improvements, including broad refactors when warranted, over bandage fixes such as tiny helper extractions or local polish.
30architecture-md-writer
Create, update, review, and split ARCHITECTURE.md files that explain a codebase's shape, major components, runtime flow, code map, and important invariants. Use when a repository lacks architecture docs, an existing ARCHITECTURE.md is stale or too detailed, a subsystem needs its own nested ARCHITECTURE.md, or a root architecture doc should link to deeper module architecture docs.
27agents-md-writer
Create, edit, review, and improve AGENTS.md files for repositories used by agentic coding tools with concise, actionable instructions and correct precedence behavior. Use whenever AGENTS.md content is being changed, including updating existing guidance, drafting a new AGENTS.md, migrating legacy instruction files, defining nested overrides in monorepos, or debugging why tools load the wrong guidance.
26source-investigator
Investigate external libraries, frameworks, and unfamiliar repositories by cloning the exact repo into a project-local temp workspace, ignoring that workspace in git, and delegating code reading to focused subagents so the main thread stays clean. Use whenever docs are incomplete, version-specific behavior matters, you need to learn how a codebase works, or exploring lots of source inline would pollute the main context.
24