Skills and agents built at Vets Who Code that help nonprofits, job seekers and small businesses create and operate straight from the terminal.
Status: three skills shipped; the content engines below are in planning. The work is tracked in issues.
| Skill | What |
|---|---|
| hashflagify | Turns a user's <login>/<login> repo or an org's .github repo into a self-updating Mission Brief profile README in VWC colors. Live on Vets-Who-Code and jeromehardaway. |
| resume-market-scorer | Scores a Software, Data, ML or AI Engineer resume 0–100 against the market for the target role: live job postings plus O*NET in-demand skills, with the BLS outlook as context. Ranks what to add, prove or cut, plus ATS wording fixes. Run it before editing a resume. |
| project-demo | Turns a portfolio repo (local path or GitHub URL) into a case study with a hero image and audio overview, plus a 60-second demo video in the builder's brand. Every feature cites a code line; README promises the code doesn't keep are called out, never repeated. |
In Claude Code:
/plugin install hashflag --marketplace Vets-Who-Code/hashflag-skills
Or copy one skill's folder into ~/.claude/skills/. Node skills need Node 22.18 or later; if you load the plugin with --plugin-dir or copy a folder, run npm ci at the repo root once.
Every content skill brings its own brand and keys: no Vets Who Code branding, keys or hosting are built in. The shared brand-pack format, paths, flags and key names are in docs/contracts.md.
The first set is content engines: one piece of writing in, images, audio and video out.
| Issue | What | Context |
|---|---|---|
| #1 | Epic: content engine skills | issue-1.md |
| #2 | Blog to image and audio | issue-2.md |
| #3 | Brand pack and brand-locked explainer video | issue-3.md |
| #4 | Project to demo video and case study for job seekers. Shipped as project-demo. | issue-4.md |
| #5 | Impact story video, audio and social images for nonprofits | issue-5.md |
| #6 | Donor update email with audio and images | issue-6.md |
| #7 | Product explainer video for small businesses | issue-7.md |
| #8 | Agent: repurpose one post into a full content kit | issue-8.md |
| #9 | Fictional sample content for testing every skill | issue-9.md |
Build order: #2, #3, #4, then #8. #5–#7 follow #2 and #3. #9 has no dependencies.
docs/context/ holds one context doc per issue. Each one covers what already exists (mostly in vets-who-code-app and the HyperFrames video pipeline), lessons already learned, costs, how each acceptance criterion maps to existing work, open decisions with recommended defaults, and a suggested build plan.
Start with issue-1.md for repo-wide decisions: layout, packaging, licensing and shared configuration.
These docs were written on 2026-09-29. Code moves, so check a file and line against the current source before relying on it.
New here? Start with #9. It needs no code.
- Branch off
mainand open a pull request. Don't commit tomaindirectly. - Use Conventional Commits:
<type>(scope): Sentence-case subject. - Run
npm testbefore opening a pull request. - Never commit licensed fonts (for example Gilroy or Gotham Pro), API keys, or rendered media output.
Code is Apache-2.0; see NOTICE. Everything in fixtures/ is CC0 1.0.