content-forge/my-commands.md

4.9 KiB
Raw Blame History

My Commands

Custom slash commands for content-forge, built from the Vin Obsidian Workflows lesson.

All command files live in .claude/commands/.

/my-world

File: .claude/commands/my-world.md

Loads full vault context in one shot. Scans all pipeline folders (00-inbox through 04-published), maps the knowledge graph (tags, orphans, dead ends, unresolved links), and summarizes priorities and connections.

When to use: Start of every session. Replaces re-explaining your project context.

/emerge

File: .claude/commands/emerge.md

Surfaces hidden patterns — ideas the vault implies but you never explicitly stated. Reads all notes, runs graph analysis (orphans, tags, dead ends), and finds emergent conclusions from scattered premises.

When to use: Weekly review, or whenever you've added several new notes and want to see what they add up to.

/challenge

File: .claude/commands/challenge.md

Pressure-tests your beliefs using the vault's own history. Finds contradictions, tensions, and counter-evidence within your notes. Forces reconciliation of conflicting ideas.

When to use: Before publishing, before major decisions, or when you suspect confirmation bias.

/connect

File: .claude/commands/connect.md

Bridges two disparate topics using the vault's link graph. Picks seemingly unrelated notes and finds the hidden bridge between them.

When to use: When looking for cross-domain insights, novel content angles, or creative connections.

/today

File: .claude/commands/today.md

晨间仪式:扫描看板状态 + 识别超期项目 + 输出今日建议。检查各阶段 WIP找出 7 天以上未更新的笔记,给出优先处理建议。

When to use: 每天开始工作时,快速了解当前状态和优先级。

/close

File: .claude/commands/close.md

晚间仪式:检查今日完成 + 更新看板状态 + Git 同步 + 更新 progress.md。确保工作有始有终。

When to use: 每天结束工作前,收尾和复盘。

/promote

File: .claude/commands/promote.md

Analyzes published content for promotion opportunities using the Promotor Blueprint framework. Scans 04-published, identifies content without promotion status, and suggests traffic sources, holding patterns, selling events, and repurpose opportunities.

Framework: Traffic → Holding Pattern → Selling Event → Loop

When to use: Weekly promotion planning, after publishing new content, when planning lead magnets or webinars.

/multi-review

File: .claude/commands/multi-review.md

Runs multi-agent content review using MCO (Multi-CLI Orchestrator). Parallel review with Claude + Codex, automatic deduplication, consensus/divergence summary.

When to use: Before publication (03-review stage), for technical accuracy checks, or when you want diverse feedback on content quality.

.

/repurpose

File: .claude/commands/repurpose.md

Transform published content into formats suitable for other platforms (X threads, LinkedIn, newsletter, video script, etc.). Based on Glue Coding principles: reuse mature content, adapt for new contexts.

When to use: After publishing (04-published stage), when planning content distribution, when you want to maximize content reach across platforms.

/trendjack

File: .claude/commands/trendjack.md

Analyze trending topics (tweets, news, events) for content creation opportunities. Uses Chain-of-Thought reasoning, Few-Shot examples, and Constitutional AI self-review to produce ROI-optimized content recommendations.

When to use: When a hot topic emerges in your field, when you want to leverage trending discussions for reach, when you need fast content ideas with built-in quality checks.

/review

File: .claude/commands/review.md

Runs structured four-dimensional content review (structure, factual, style, format) on drafts in 03-review/. Outputs a pass/fail decision with issue tracking, updates frontmatter fields (review_status, review_issues, review_passed_at), and manages revision workflow.

Review Dimensions:

Dimension Focus Critical Issues
Structure Logic flow, paragraph structure, info hierarchy Missing sections, broken logic
Factual Data accuracy, references, technical correctness Fabricated data, broken sources
Style Audience match, voice consistency, readability Wrong tone, jargon misuse
Format Markdown compliance, frontmatter, asset refs Invalid YAML, missing fields

Pass Criteria: Critical = 0 AND High ≤ 1

Revision Rules:

  • revision_count ≤ 3 per article
  • Overflow → blocked for human decision
  • After revision → returns to review

When to use: After moving draft to 03-review/, before polish phase, when quality gate is needed before publication.

Integration: Works with /multi-review for multi-agent review and /challenge for pressure-testing key claims.