Instagram-agent-skill
AGENTS.md1 starsChanged 11 days agoMITView on GitHub
This project contains 13 agent skills and six local Python command-line tools. The tools use only the Python standard library; use Python 3.10 or newer. The Python tools need no application server or API key. The web/ directory contains a standalone browser application; see WEBSITE_RU.md. Canonical skill folders live in skills/ig-*. Project-local discovery folders .agents/skills/ (Codex) and .claude/skills/ (Claude Code) link to those same folders. Read the relevant SKILL.md before using a skill. Run a Python script from its own…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 32,671 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository13 other files it gives its agents.
- ig-auditSkill
- ig-captionSkill
- ig-carouselSkill
- ig-commentSkill
- ig-dmSkill
- ig-humanSkill
- ig-planSkill
- ig-profileSkill
- ig-reelSkill
- ig-replySkill
- ig-repurposeSkill
- ig-storySkill
- ig-viralSkill
The file
We split this AGENTS.md into the rules it gives an agent
7 rules. About 1 is common, 1 is distinctive, and 5 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 37
Run python3 -m unittest discover -s tests -v after changes.
- Very specific: line 45
Keep user input escaped when rendering HTML and preserve the language limits.
- Very specific: line 46
Do not describe prompt assembly as AI generation or imply Instagram is connected.
- Very specific: line 49
Run python3 scripts/check_navigation_browser.py after navigation/startup changes.
- Common: line 57
Never put Meta secrets or Instagram tokens in frontend files, Git, or logs.
- Distinctive: line 58
Do not claim a live connection based on emulator tests.
- Very specific: line 65
Keep provider credentials in secure settings.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md 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
Instagram agent skills
This project contains 13 agent skills and six local Python command-line tools.
The tools use only the Python standard library; use Python 3.10 or newer.
The Python tools need no application server or API key. The web/ directory
contains a standalone browser application; see WEBSITE_RU.md.
Skills and scripts
Canonical skill folders live in skills/ig-*. Project-local discovery folders
.agents/skills/ (Codex) and .claude/skills/ (Claude Code) link to those same
folders. Read the relevant SKILL.md before using a skill. Run a Python script
from its own skill folder, or give its path from the project root, for example:
python3 skills/ig-reel/hookscore.py --hook 'I lost $18,000 on one missing contract.'
python3 -m unittest discover -s tests -vProject-local account data
For this project, interpret the upstream skills' ~/.claude/instagram/ paths
as .instagram/ under this checkout. This applies to voice.md, swipe.md,
plan.md, and log.md. Keep account data here rather than writing global
configuration. The directory is ignored by Git. If voice.md is missing,
copy templates/voice.md there without overwriting any existing profile.
An empty template is not a completed voice profile; ask for the user's real
examples and details when a content task requires them.
The lexicon, hook scoring, speech timing, and text patterns are designed for English. Do not present scores or timings for Russian or other languages as validated. Follow the user's language preference when writing content and explain the tools' language limits when relevant.
Validation and provenance
Run python3 -m unittest discover -s tests -v after changes. It exercises all
six CLIs, their positive and negative paths, JSON assets, and skill links.
See SETUP_RU.md for local usage and the upstream revision. Preserve the MIT
license and author attribution.
For website changes, run python3 scripts/build_site.py to refresh web/data.js
and the standalone preview in dist/. JavaScript uses canonical constants
exported from the original Python tools. The website is static and works offline.
Keep user input escaped when rendering HTML and preserve the language limits.
Do not describe prompt assembly as AI generation or imply Instagram is connected.
web/studio.js is the generated, versioned browser bundle. Edit its original
source files and rebuild; the published HTML must load one coherent bundle.
Run python3 scripts/check_navigation_browser.py after navigation/startup changes.
It covers real clicks, mobile menus, sandboxed previews, browser history and
optional scripts failing to load. These checks also run before Pages deployment.
The optional Instagram integration lives in backend/ (Cloudflare Workers/D1)
and web/instagram.js. Use Node.js 22+, npm ci --prefix backend,
npm test --prefix backend, and npm run check --prefix backend for its checks.
See INSTAGRAM_SETUP_RU.md for real Meta application and hosting requirements.
Never put Meta secrets or Instagram tokens in frontend files, Git, or logs.
Do not claim a live connection based on emulator tests. Every publication needs
explicit confirmation of its media, caption, and account.
The optional Apify bridge is backend/src/apify.mjs; an independent Worker can
use backend/wrangler.apify.toml without Meta or D1. web/apify-data.js is the
shared allowlisted projection for browser imports, backend responses, and the
collection CLI. Run python3 scripts/check_apify_browser.py for its UI checks.
Keep provider credentials in secure settings; the site accepts only a distinct
studio access code. Do not label a caption score as a reviewed video hook.
See APIFY_SETUP_RU.md; real collection needs a configured APIFY_API_TOKEN.
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.

