From Recharm
Show a summary table that shows the search queries for every scene and the number of clips you evaluated before making the final recommendation in the brief.
Configuration options for this output style
This output style defines how the brief-creator skill presents results to the user. Follow this format precisely after completing all searches and clip selections. The goal is full transparency — the user should be able to see exactly what was searched, what was found, and why clips were chosen, without needing to re-open Recharm.
Begin with a concise summary of the brief and what was accomplished.
# 🎬 [Brief Title]
**Concept:** [One sentence summary of the creative direction]
**Sections completed:** [N]
**Total clips selected:** [N]
**Searches performed:** [N]
For each section of the brief, output a dedicated block in this format:
## Section [N]: [Section Name]
**Creative direction:** [What this section needs to convey — pulled from the brief]
### 🔍 Searches Performed
| # | Query | Filters Applied | Clips Found | Clips Chosen |
|---|-------|----------------|-------------|--------------|
| 1 | "[search query]" | Label: [label], Duration: [range], etc. | [N] | [N] |
| 2 | "[search query]" | Label: [label] | [N] | [N] |
> **Why these queries?** [1–2 sentences explaining the search strategy for this section — what visual idea you were looking for and why you approached it this way.]
### 🎞️ Selected Clips
For each chosen clip:
**Clip [N]** — [clip title or ID]
- **Why selected:** [What made this clip the right fit — composition, motion, mood, subject match]
- **Suggested timing:** [Where in the section this clip works best, e.g. "opening shot", "cutaway", "closing hold"]
- [Recharm link or thumbnail if available]
> **What was skipped and why:** [Briefly note if many clips were found but few chosen — explain the selection criteria used, e.g. "Filtered out clips with logos visible" or "Preferred wide shots over close-ups for this section."]
Repeat this block for every section in the brief.
After all sections, output a compact summary table the user can scan at a glance:
## 📋 Full Brief at a Glance
| Section | Clips Selected | Primary Query | Key Filter |
|---------|---------------|---------------|------------|
| [Name] | [N] | "[query]" | [filter] |
| [Name] | [N] | "[query]" | [filter] |
Always end with a recommendations block. This is not optional — it helps the user improve the brief before going into production.
## 💡 Recommendations to Strengthen This Brief
### Add more coverage for these sections
[List any sections where fewer than 2 clips were found, or where the search results were a weak match. Suggest alternative search angles or different filter combinations to try.]
### Queries to try next
- "[alternative query 1]" — [why this might surface better results]
- "[alternative query 2]" — [what different angle this covers]
### Consider broadening these filters
[If filters were narrow and resulted in low clip counts, name the specific filter and suggest loosening it, e.g. "The 'Duration: 5–10s' filter on Section 2 returned only 3 clips — try 3–15s for more options."]
### Gaps in the brief
[Note any sections of the creative concept that have no strong clip match in the library — flag these so the user knows to shoot new footage or source externally.]
### Production notes
- [Any observations about clip consistency across sections — e.g. color grading, aspect ratio mismatches]
- [Suggestions on sequencing — if a selected clip would work better in a different section]
- [Pacing notes — if too many similar clips were chosen back-to-back]
npx claudepluginhub recharm/claude-plugin --plugin recharmFormats Claude's responses for the Harness Plan/Work/Review workflow. Planning phase uses task tables with status markers; implementation shows code changes with context; review structures findings by severity. Includes progress reporting (Done/Current/Next). Preserves coding behavior.
Formats Claude's responses to be concise, native, and execution-first. Avoids ceremony, planning, or multiple options unless explicitly requested. Preserves coding behavior. Prioritizes native tools and agent/plan usage. Table-friendly output.
Enforces fact-grounded FinOps analysis: evidence-backed claims with citations/derivations, standardized currency/percentage/variance formatting, structured tables for comparisons, explicit time periods and assumption labels for cloud cost data. Preserves coding instructions.