Overcast skill creator

Skill kdr/overcast/skills/overcast-skill-creator

Create small, installable agent skills that wrap focused Overcast workflows. Use when the user asks to make an Overcast skill for a specific investigation, media analysis, recon, monitoring, or case-memory workflow.From its SKILL.md

Install

npx -y skills add kdr/overcast --skill overcast-skill-creator

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

12 stars12 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.8 KB, 636 tokens by cl100k_base, as published. Nobody here has run it

overcast-skill-creator

Use this when the user wants a focused skill built on Overcast instead of the broad overcast skill. Example requests: "make an Overcast skill for analyzing security camera clips", "create a skill that monitors a target and briefs me", or "turn this Overcast workflow into an installable agent skill".

Reference the broad overcast skill and its overcast/reference/verbs.md man pages for exact flags. Do not duplicate the full verb reference.

Design Rules

Pick one case lifecycle: initialize/setup, gather or sense evidence, add notes/findings, ask/brief, then export.
Choose the minimum verbs needed. Prefer case setup, watch, listen, see, face, scan, capture, monitor, note, finding, ask, and brief only when they serve the workflow.
Preserve citations. Evidence claims should cite record.id plus media.at when a timestamp or range exists.
Prefer ask and brief over raw JSON spelunking for synthesis. Use raw records for verification and exact fields, not as the default reading path.
For large watch content or listen transcripts, use case memory get <record-id> --field <field> --offset <n> --limit <n> rather than head/tail reads of JSONL.
State setup assumptions: overcast doctor, provider credentials, system ffmpeg/ffprobe, tinycloud version, and whether the workflow needs live sources or only local files.

Template

---
name: overcast-<workflow-name>
description: >-
  <One sentence about when an agent should use this focused Overcast workflow.>
---

# overcast-<workflow-name>

Use this skill when <trigger conditions>.

## Quickstart

```bash
overcast doctor --json
overcast case init --json
overcast case setup --target "<target>" --yes --json
overcast <gather-or-sense-verb> <input> --json
overcast ask "<question>" --json
overcast brief --export ./brief.md --json
```

## Evidence Rules

- Cite `record.id` and `media.at` for every media-derived claim.
- Record human observations with `note --ref <record-id> --at <time-range>`.
- Separate observed facts, inferred expected behavior, and open questions.

## Failure Handling

- Run `overcast doctor --json` when a provider or system dependency fails.
- If a record field is large, page it with `case memory get`.
- If a source is unavailable, report the missing source and continue with local
  case evidence.

## Validation

```bash
overcast commands --json
overcast <main-verb> --help
overcast ask "<workflow-specific verification question>" --json
```

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most skill authoring skills give in 636 tokens

Counted across 408 of the 422 authors here whose files we hold, read 2026-09-06

Keep SKILL.md under 500 linesin 118 of 408, across 104 files
Write skill instructions in imperative formin 44 of 408, across 37 files
Move detailed reference material into references filesin 41 of 408, across 34 files
Run init_skill.py to initialize new skillsin 40 of 408, across 33 files
Draft assertions while test runs are in progressin 32 of 408, across 25 files
Keep description under 1024 charactersin 31 of 408, across 18 files
Run package_skill.py once the skill is completein 29 of 408, across 23 files
Spawn with-skill and baseline runs in the same turnin 28 of 408, across 21 files
Test added scripts by actually running themin 27 of 408, across 21 files
Follow the creation steps in orderin 24 of 408, across 17 files
Generate the eval viewer before evaluating outputs yourselfin 24 of 408, across 17 files
Write the description in third personin 24 of 408, across 18 files

Said here and by no other author read

Pick one case lifecycle
Choose the minimum verbs needed
Cite record.id and media.at for evidence claims
Prefer ask and brief over raw JSON
Page large records with case memory get
State setup assumptions including dependencies and credentials

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.