Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/ohswedd/praxis/capability-discoverynpx skills add Ohswedd/praxis --skill capability-discoverygit clone --depth 1 https://github.com/Ohswedd/praxisWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ohswedd/praxis/capability-discovery)<a href="https://agentmods.dev/skills/ohswedd/praxis/capability-discovery"><img src="https://agentmods.dev/badge/skills/ohswedd/praxis/capability-discovery.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00109 | $0.00833 |
| Opus 5 | $0.00055 | $0.00417 |
| Sonnet 5 | $0.00022 | $0.00167 |
| Haiku 4.5 | $0.00011 | $0.00083 |
Grade A, and why
capability-discovery scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Capability Discovery
Close the gap between "this task needs X" and "I have X", without reinventing what already exists. Claude Code does not auto-install capabilities, so this skill provides the disciplined workflow to do it deliberately.
Step 1: Name the gap precisely
State exactly what is missing and why the task needs it, e.g. "needs to query the project's Postgres to validate the migration", "needs to drive the browser to verify the UI change", "this three-step release check recurs and should be a command".
Step 2: Search existing first (no reinvention)
In order, check:
- Already available: is there a configured MCP server or an installed
plugin/skill that covers it? (
/mcp,/plugin list, the available tools.) - In a marketplace: is there a maintained plugin/MCP for it? Prefer a well-maintained existing one over building your own.
- In the repo: is there already a script or Make target that does this?
Only proceed to create something if all three come up empty.
Step 3: Choose the smallest right primitive
- Recurring prompt → a slash command.
- Real domain logic or a multi-step workflow with helper files → a skill.
- Isolated, parallel, or read-limited work → a subagent.
- Access to an external system over a protocol → an MCP server (reference credentials via environment variables; never hard-code secrets).
- A rule to enforce deterministically → a hook.
Step 4: Scaffold ad-hoc, cleanly
Create the minimal working version in .claude/ (fast to iterate). Give it a
crisp, "pushy" description so it triggers when relevant. Keep it small, every
component adds always-on context. Document how to use it.
Check the workspace mode before you write it
(python3 "${CLAUDE_PLUGIN_ROOT}/scripts/config.py" status). A capability is a
file in the project's tree like any other, and in contributor mode the project
did not ask for it:
- Scaffold under
.claude/.praxis/capabilities/instead of.claude/, which is inside the git-excluded state directory, and load it explicitly rather than by convention. A command or skill you invented for one task is not a contribution. - Put MCP servers in
.claude/settings.local.json(or user scope), never in a committed.mcp.json. That file exists to be shared with the team, and the team here is not yours. - If the capability is genuinely something the project needs, that is a proposal for its maintainers: raise it in the pull request or an issue, with the reasoning, rather than committing it alongside an unrelated change.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 65 lines · 109 tokens per session scan A 60358de27102
capability-discovery is a skill published in the GitHub repository Ohswedd/praxis (1 stars, last pushed 1mo ago), licensed MIT. It adds 109 tokens to every session and 833 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
scroll-craft
Build premium scroll-driven landing pages for service, product, food, and drink brands. Plan the visitor journey, page grammar, emotional peak, and bespoke signature move. Create dimensional heroes with independent visual planes, restrained motion, and separate mobile composition. Use supplied photos and footage or…
a11y-audit
Dedicated WCAG 2.2 AA/AAA accessibility audit across 10 dimensions (A1-A10) covering semantic HTML, keyboard navigation, ARIA patterns, color contrast, forms, images/media, responsive/zoom, motion/animation, reading/content, and legal compliance. Goes far beyond surface-level design-review checks with deep…
wp-demo
Demo HTML creation methodology — single-file demos with section comments for 1:1 WordPress conversion.
wp-polylang
Polylang multilingual methodology — one post per language joined by translation groups, driven through the pll API.
wp-responsive
Responsive design patterns — mobile-first breakpoints, fluid typography, responsive images, touch targets.
wp-cli-patterns
WP-CLI-first principle — use WP-CLI instead of generating PHP code whenever possible, with command reference, ACF seeding patterns, and environment-aware execution.