Write or rewrite prose so it sounds like a human wrote it — not a frontier model. Named in homage to Rabindranath Tagore, whose prose carried what frontier models reach for and miss: a point of view, specificity over abstraction, and restraint over puffery. Merges two complementary approaches: a 29-pattern catalog of AI tells (from humanizer) plus an 8-rule operating system with an 8-dimension scoring gate (extending stop-slop). Use when drafting, editing, or reviewing any prose: essays, posts, docs, reports, emails. Detects and removes inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary, passive voice, negative parallelisms, filler phrases, inanimate-verb constructions, narrator-from-a-distance voice, and metronomic rhythm. Adds back the things AI writing usually lacks: point of view, stakes, specificity, restraint, varied rhythm, and trust in the reader.
$npx -y skills add apurvrdx1/tagore --skill tagoreInstalls into the current project.
Run `npx skills use "https://github.com/apurvrdx1/tagore" --skill "apurvrdx1/tagore"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/apurvrdx1/tagore" that are relevant to the current task. Run `npx skills add "https://github.com/apurvrdx1/tagore"` and select the relevant skills, then follow their instructions.
| 1 | # Tagore |
| 2 | |
| 3 |  |
| 4 | |
| 5 | > *"The butterfly counts not months but moments, and has time enough."* |
| 6 | > — Rabindranath Tagore |
| 7 | |
| 8 | A skill for Claude Code, OpenCode, GitHub Copilot CLI, OpenAI Codex CLI, Gemini CLI, Goose, and Hermes (NousResearch agent runtime). Run `/tagore` on the AI-drafted prose you just generated; it rewrites the text to sound human and scores how close it got. Named in homage to Rabindranath Tagore, whose prose carried what frontier models reach for and miss: a point of view, specificity over abstraction, and restraint over puffery. |
| 9 | |
| 10 | [](https://www.buymeacoffee.com/apurvrdx) |
| 11 | |
| 12 | ## What it does |
| 13 | |
| 14 | Tagore attacks slop in two directions at once: |
| 15 | |
| 16 | - **Removes the tells.** 29 cataloged AI patterns get scrubbed: significance inflation, em dash overuse, vague attributions, AI vocabulary, rule of three, copula avoidance, and more. |
| 17 | - **Adds what's missing.** Point of view, stakes, specificity, restraint, varied rhythm, and trust in the reader. |
| 18 | |
| 19 | The skill runs every rewrite through a six-stage pipeline ending in an 8-dimension scoring rubric (5 mechanics + 3 substance) with a 56/80 pass threshold and subtotal floors so a piece can't pass on mechanics alone. |
| 20 | |
| 21 | ## Why merge two skills |
| 22 | |
| 23 | Tagore is a synthesis of two complementary skills: |
| 24 | |
| 25 | | Source | Contribution | |
| 26 | |--------|--------------| |
| 27 | | [humanizer](https://github.com/blader/humanizer) by [blader](https://github.com/blader) (based on Wikipedia "Signs of AI writing") | 29-pattern catalog, voice calibration, personality/soul section, self-audit loop, full worked example | |
| 28 | | [stop-slop](https://github.com/hardikpandya/stop-slop) by [Hardik Pandya](https://hvpandya.com) | 8 core principles, 12-item pre-delivery checklist, scoring rubric | |
| 29 | |
| 30 | humanizer is a deep catalog with a pipeline. stop-slop is a tight rulebook with a scoring gate. Tagore fuses them into a single workflow and adds three substance dimensions (Specificity, Restraint, Voice) that neither skill scored on its own. |
| 31 | |
| 32 | ## Install |
| 33 | |
| 34 | ### One-liner (auto-detect) |
| 35 | |
| 36 | ```bash |
| 37 | curl -fsSL https://raw.githubusercontent.com/apurvrdx1/tagore/main/install.sh | bash |
| 38 | ``` |
| 39 | |
| 40 | The installer detects every supported harness on your system and installs Tagore into each. It also drops a universal symlink at `~/.agents/skills/tagore/` so any agent honoring the cross-tool convention can find it. |
| 41 | |
| 42 | ### Install for one harness only |
| 43 | |
| 44 | ```bash |
| 45 | curl -fsSL https://raw.githubusercontent.com/apurvrdx1/tagore/main/install.sh | bash -s -- --platform claude |
| 46 | ``` |
| 47 | |
| 48 | Supported `--platform` values: `claude`, `opencode`, `copilot`, `codex`, `gemini`, `goose`, `hermes`, `agents` (universal). |
| 49 | |
| 50 | ### Install for every detected harness |
| 51 | |
| 52 | ```bash |
| 53 | git clone https://github.com/apurvrdx1/tagore.git && cd tagore && ./install.sh --all |
| 54 | ``` |
| 55 | |
| 56 | ### Manual install |
| 57 | |
| 58 | Pick your harness and copy the `tagore/` directory (with `SKILL.md` inside) into the matching location: |
| 59 | |
| 60 | | Harness | Global install path | Per-project alternative | |
| 61 | |---------|---------------------|-------------------------| |
| 62 | | Claude Code | `~/.claude/skills/tagore/` | `.claude/skills/tagore/` | |
| 63 | | OpenCode | `~/.config/opencode/skills/tagore/` | `.opencode/skills/tagore/` | |
| 64 | | GitHub Copilot CLI | `~/.copilot/skills/tagore/` | `.github/skills/tagore/` | |
| 65 | | OpenAI Codex CLI | `~/.codex/skills/tagore/` | `.agents/skills/tagore/` | |
| 66 | | Gemini CLI | `~/.gemini/skills/tagore/` | — | |
| 67 | | Goose | `~/.config/goose/skills/tagore/` | — | |
| 68 | | Hermes (NousResearch) | `~/.hermes/skills/writing/tagore/` (or `$HERMES_HOME/skills/writing/tagore/`) | `.hermes/skills/writing/tagore/` | |
| 69 | | Cursor | — | `.cursor/rules/tagore/` | |
| 70 | | Windsurf | — | `.windsurf/rules/tagore/` | |
| 71 | | Universal (cross-tool) | `~/.agents/skills/tagore/` | `.agents/skills/tagore/` | |
| 72 | |
| 73 | The universal `~/.agents/skills/` path is recognized by OpenCode, Copilot CLI, |