github-raw-reader
Skill0 starsChanged 3 months agoMITView on GitHub
Teach agent access public repo files via GitHub Raw URLs. Bypass HTML pages for direct source code parsing.
In this repository24 other files it gives its agents.
- corporate-espionageSkill
- dispatching-parallel-agentsSkill
- epub-readerSkill
- executing-plansSkill
- finishing-a-development-branchSkill
- magisk-moduleSkill
- notepadpp-plugin-skillSkill
- planningSkill
- receiving-code-reviewSkill
- requesting-code-reviewSkill
- reviewSkill
- skill-creator2Skill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- technical-writerSkill
- test-driven-developmentSkill
- text-humanizerSkill
- using-git-worktreesSkill
- using-superpowersSkill
- verification-before-completionSkill
- windows-diagnosticsSkill
- writing-plansSkill
- zed-extensionSkill
- zed-themeSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 55
Use when raw.githubusercontent.com is blocked but github.com git smart HTTP is reachable.
- Very specific: line 82
Branch name: default is usually main or master.
- Distinctive: line 83
Always fetch raw content. Do not parse rendered HTML for code.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
GitHub Raw Access Skill
Access repository files via direct raw URLs or git clone. Two methods are available — use Method A first, fall back to Method B if raw.githubusercontent.com is blocked.
Method A — Raw URL (preferred)
Step 1: Enumerate files
Tree pages (/tree/) return full HTML via curl and contain all file/subdir paths. Use this to enumerate before constructing raw URLs.
Root listing:
curl -s "https://github.com/<user>/<repo>/tree/<branch>" \
| grep -oP '"/<user>/<repo>/(?:blob|tree)/<branch>/[^"&]+"' \
| sort -uSubdir listing (recurse for each tree/ path found):
curl -s "https://github.com/<user>/<repo>/tree/<branch>/<subdir>" \
| grep -oP '"/<user>/<repo>/(?:blob|tree)/<branch>/[^"&]+"' \
| sort -ublob/ entries are files. tree/ entries are subdirectories — recurse into them.
Step 2: Convert blob URLs to raw URLs
| From | To |
https://github.com/ | https://raw.githubusercontent.com/ |
/blob/ | / (remove) |
Example:
- Input:
https://github.com/14ag/python-arc/blob/master/2024/snake_essentials.py - Output:
https://raw.githubusercontent.com/14ag/python-arc/master/2024/snake_essentials.py
Step 3: Fetch file content
curl -s "https://raw.githubusercontent.com/<user>/<repo>/<branch>/<path/to/file>"Method B — Git sparse clone (fallback)
Use when raw.githubusercontent.com is blocked but github.com git smart HTTP is reachable. This uses the git pack protocol over HTTPS, which is allowed even in restricted environments.
# Clone with no blobs (tree only), then sparse-checkout the needed paths
git clone --depth=1 --no-tags --filter=blob:none --sparse \
https://github.com/<user>/<repo>.git /tmp/repo
cd /tmp/repo
git sparse-checkout set <subdir1>/ <subdir2>/
# Read any file directly
cat /tmp/repo/<path/to/file>To verify git smart HTTP is reachable before cloning:
curl -s "https://github.com/<user>/<repo>.git/info/refs?service=git-upload-pack" | head -c 4
# Should return a git pkt-line (e.g. "001e"), not "Host not in allowlist"Notes
/tree/pages: accessible viacurl, return full HTML with file listings./blob/pages: return 0 bytes viacurl(JS-rendered). Do not use for enumeration.github.com/raw/<branch>/<path>: redirects toraw.githubusercontent.com— blocked if the CDN domain is.- Branch name: default is usually
mainormaster. Checkpacked-refsafter clone, or inspect the root tree page title if unsure. - Always fetch raw content. Do not parse rendered HTML for code.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

