.md file to compare - side-by-side diff against snappy-https
snappy-https
What it does for you
Lets your assistant open links, images, and files you share into chat.
What it produces
A recent result, so you can see the kind of work it returns.
loading…
How to get it
These run inside the Snappy workspace. Want this working in your business? I set skills like this up with you, in one focused week.
For developers how this skill is built, graded, and how it runs
at a glance- the short version
what's inside - the parts that make up a skill 3/4 present
A skill is just a few plain-text files. Only the main one is required. The rest are optional, added as the work needs them. This is what the skill is made of; how it runs is just below.
state/skills/snappy-https/SKILL.md
present
state/skills/snappy-https/api.ts
present
state/bin/snappy-https/
not present
state/skills/snappy-https/AGENTS.md
present
how it runs - the shared frame every skill uses 5/5 present
Every skill runs the same way. One part does the work, a separate part checks it, and a short loader hands the AI exactly what it needs for the job. Anything this skill doesn't use shows a one-line note saying why, on purpose, not by accident.
state/log/evals.ndjson what it has learned - fixes written back in over time sample
When a run hits something this skill didn't handle, the fix gets written back into the skill so it doesn't happen again. FIXED means it was corrected on the spot. LOGGED means it's queued for a bigger rewrite. Either way, the skill gets a little better and never makes the same mistake twice.
- Loading feedback rows…
how the work flows- who makes it, who checks it
what this step does
what this step does
SKILL.md- the skill, written out in plain English
snappy-https
Handles URLs shared in chat that point to external resources - DigitalOcean Spaces objects, images, zips, or any HTTPS link. Prevents the failure mode where the user pastes a URL expecting the agent to fetch or inspect it, and the agent ignores the link entirely.
Observed user requests
These are the prompts that triggered this skill being scaffolded:
- "https://robert-storage.tor1.digitaloceanspaces.com/shared/2026-04-28-1777395923733-crayon-sdk.zip -- Okay, so I just now"
- "https://robert-storage.tor1.digitaloceanspaces.com/shared/2026-04-28-1777409614425-snappy-region-1777409610914.png -- sc"
- "https://robert-storage.tor1.digitaloceanspaces.com/shared/2026-04-29-1777436955168-snappy-region-1777436950892.png -- th"
Steps
- Detect the URL's content type via a HEAD request (
curl -sI <url>) - checkContent-Typeresponse header. - For image types (
image/*): download to/tmp/snappy-fetch-<timestamp>.<ext>and display viastate/lib/agent-browser.ts screenshotor pass tostate/lib/image.tsfor analysis. - For archive types (
application/zip,application/x-tar, etc.): download to/tmp/, extract withunzip -lortar -tzfto preview contents without full extraction. - For text/JSON types:
curl -sL <url>and surface the first 200 lines. - For DigitalOcean Spaces URLs (
*.digitaloceanspaces.com): no auth required for public objects; private objects need DO Spaces credentials viaenv("DO_SPACES_KEY"). - Log the fetch result (url, content-type, byte-size, disposition) as an eval row via
score("snappy-https", run_id, { score: 1.0 }). - Never store fetched content in git or
.env.cache- temp files under/tmp/only.
Steps (scope-only / apply:false)
1. Scope - no side effects
What to read, what to compute, what to return. This step MUST be runnable with no apply: true flag and no credentials beyond read-only.
2. Gate
Hard-fail any missing required field. Never soft-skip. The gate is the protection against "I thought we checked that."
3. Act (only if the gate passes)
The real work. Dispatch, write, post, or compose. Keep this step thin - heavy logic belongs in state/lib/snappy-https.ts (or state/bin/snappy-https/), not inline here.
4. Log + eval
import { score } from "../../lib/eval";
score("snappy-https", run_id, {
score: <pass = 1.0 | partial = 0.5 | fail = 0.0>,
primary_issue: <null | one-line reason>,
});
Eval
Actor: the thing that produces the output (a dispatch model, a CLI, or the state/lib/snappy-https.ts library if one exists). Auditor: the thing that judges (must be different - see CONSTITUTION invariant #3). Name both explicitly.
Score convention:
| Outcome | Score |
|---|---|
| Pass on first try | 1.0 |
| Failed first, auto-fix applied, re-check passed | 0.5 |
| Still failing or unrecoverable | 0.0 |
If you cannot name a deterministic auditor, switch the frontmatter to eval: manual and log to state/log/pending-eval.ndjson - but fight to avoid manual. Manual is the escape hatch that leaks the thesis.
Gotchas
- List concrete failure modes you hit while building. Examples: "API
returns 200 with empty body"; "round-tripped text has a , artifact where em-dashes used to be"; "cache file is stale after 12h."
- If a gotcha matches an entry in user memory (see
~/.claude/CLAUDE.md),
cite the memory name so future agents can follow the trail.
AGENTS.md- what the AI loads when this skill comes up
snappy-https - loader
Per-turn rules for snappy-https. Full reference: state/skills/snappy-https/SKILL.md.
Critical Rules
- Default scope-only (apply:false). Preview the download; require explicit confirmation before any write.
- HEAD probe first.
curl -sI <url>extractsContent-Typeheader. Route by type. - DO Spaces public (no auth by default). Private objects need
env("DO_SPACES_KEY"). - Never persist to git. Temp files
/tmp/snappy-fetch-<timestamp>.<ext>only. - Eval row mandatory. After run:
score("snappy-https", run_id, { score: 1.0 }).
Commands
| ui dashboard | state/skills/snappy-https/resources/ui.openui |
| operation | command | |
|---|---|---|
| HEAD probe | `curl -sI <url> \ | grep Content-Type` |
| download to /tmp | curl -sL <url> -o /tmp/snappy-fetch-$(date +%s).<ext> | |
| image analysis | pass to state/lib/image.ts or display via agent-browser | |
| archive preview | unzip -l <file.zip> or tar -tzf <file.tar.gz> | |
| text first 200 | `curl -sL <url> \ | head -200` |
| eval log | state/log/evals.ndjson | |
| writeback | state/log/loader-feedback.log |
Self-Test
An agent reading this should correctly:
- [ ] Detect content-type via HEAD before downloading?
- [ ] Route image/zip/text by type?
- [ ] Use DO Spaces public access without explicit auth?
- [ ] Write eval row with score=1.0 on success?
- [ ] Never commit temp files to git?
- [ ] Pass DO Spaces URLs (*.digitaloceanspaces.com) as no-auth by default?
Self-correcting loader (PID feedback)
If this loader didn't cover your case - missing content type, no DO Spaces fallback, wrong temp path - you MUST attempt an Edit to this AGENTS.md before you log.
Log before exit:
echo "[$(date -u +%FT%TZ)] snappy-https: <what was missing or fixed> [FIXED|LOGGED] action_kind=skill-ran" >> state/log/loader-feedback.log
FIXED= you edited this file inline.LOGGED= the fix was too large; it's queued for regen.
OpenUI Resource
- Skill-owned OpenUI Lang resource:
state/skills/snappy-https/resources/ui.openui. Read it before rendering or editing this skill's generated component surface. - Treat this resource as a first-class artifact of the skill, not a generic chat response. Improve it when the skill's user-facing output needs to become richer.
- System resources compose OpenUI primitives and inherit SnappyChat tokens. Use
ui_contract: brandedin SKILL.md only for deliberate platform or client visuals.
api.ts- the code it can call
#!/usr/bin/env npx tsx
/**
* state/skills/snappy-https/api.ts — sidecar stub for the snappy-https skill.
*
* This file is created by the scaffolder so a fresh skill folder is
* structurally valid. Replace the placeholder with the real implementation
* the moment the skill needs executable logic, OR move the implementation
* to `state/lib/snappy-https.ts` (preferred — the lib path is what
* `eval: shape` validates against).
*
* If this skill has no backing code (prose-only slash command), delete this
* file and rely on `eval: auto-shape` in SKILL.md.
*/
export const SKILL_NAME = "snappy-https" as const;
export function describe(): string {
return "The https skill — purpose TBD.";
}
if ((() => { try { return import.meta.url === `file://${process.argv[1]}`; } catch { return false; } })()) {
console.log(JSON.stringify({ skill: SKILL_NAME, describe: describe() }, null, 2));
}
scripts- helper scripts it can run
prose-only skill - 1 inline code block live in SKILL.md above (no state/bin/ sidecar yet).
how we check it- the checks, plus the last 10 runs
no recent runs logged - the eval contract is declared but nothing has been graded yet