Planning With Files
jarrodwatts/claude-code-config
Transforms workflow to use Manus-style persistent markdown files for planning, progress tracking, and knowledge storage.
Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups.
Install Conventions AI
$ npx skills add stella/stella --skill conventions-ai -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install stella/stella conventions-ai --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/conventions-ai .claude/skills/conventions-ai && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .claude/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/stella/stella/tree/main/.agents/skills/conventions-aiType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add stella/stella --skill conventions-ai -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install stella/stella conventions-ai --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/conventions-ai .agents/skills/conventions-ai && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .agents/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add stella/stella --skill conventions-ai -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install stella/stella conventions-ai --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/conventions-ai .cursor/skills/conventions-ai && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .cursor/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/stella/stella.git --path .agents/skills/conventions-ai--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add stella/stella --skill conventions-ai -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install stella/stella conventions-ai --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/conventions-ai .gemini/skills/conventions-ai && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .gemini/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install stella/stella conventions-aiInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add stella/stella --skill conventions-ai -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/conventions-ai .github/skills/conventions-ai && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .github/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add stella/stella --skill conventions-ai -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install stella/stella conventions-ai --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/stella/stella.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/conventions-ai .opencode/skills/conventions-ai && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "conventions-ai" agent skill from https://github.com/stella/stella/tree/main/.agents/skills/conventions-ai into .opencode/skills/conventions-ai/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "conventions-ai", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
conventions-ai- GitHub stars
- 258
- Token cost
- ~1.6k tokens
- SKILL.md length
- 828 words
- Files
- 2
- Skills in repo
- 24
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups.
- Works in 7 steps: Every touched SDK union is exhaustive at… → Every external boundary validates… → Live, persisted, and reloaded message… → …
- Tasks that involve Structured output and tool calling
- SKILL.md covers Read the contract first, Make SDK states exhaustive, Preserve lifecycle semantics and Advertise only executable…, plus 2 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Conventions AI is an agent skill from stella/stella. Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups. Enforces exhaustive SDK state handling, truthful capability exposure, recoverable failures, and stream-to-reload parity.
Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).
It sits in AI & LLM Engineering, covering Structured output and tool calling. The repository describes itself as: Open-source legal workspace. The licence is Apache-2.0.
When your agent uses it
- Tasks that involve Structured output and tool calling
Example prompts
- “/conventions-ai”
Workflow steps
7 steps, taken from the first numbered list in SKILL.md.
- Every touched SDK union is exhaustive at compile time.
- Every external boundary validates runtime data without dropping valid states.
- Live, persisted, and reloaded message parts have the same meaning.
- Advertised tools are executable in the current context.
- Recoverable failures let the model continue without UX drama.
- Pending human interactions retain turn ownership.
- Focused tests reproduce lifecycle ordering, not only helper behavior.
What it can do on your machine
Read from SKILL.md and the folder at commit b225fd8. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Conventions AI loads about 1.6k tokens when it runs. Until then it costs about 79 tokens; SKILL.md has 828 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from stella/stella at commit b225fd8, republished under its Apache-2.0 licence (© stella). 828 words, ~1,553 tokens.
Download SKILL.mdSave it as
.claude/skills/conventions-ai/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.- name
- conventions-ai
- description
- Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups. Enforces exhaustive SDK state handling, truthful capability exposure, recoverable failures, and stream-to-reload parity.
AI Conventions
AI integrations are protocol boundaries. Keep SDK evolution, provider iteration, tool failures, persistence, and UI reconstruction from producing silent state loss.
Read the contract first
- Fetch current documentation for the exact installed AI SDK and adapters before changing lifecycle or state handling. Inspect installed types and implementation when documentation does not define event ordering precisely.
- Trace the complete vertical path: provider events, server stream bridge, client stream processor, message-part types, persistence validation, reload, and UI. A fix at one layer is incomplete if another layer can discard the same state.
Make SDK states exhaustive
- Never reinterpret or manually redeclare a protocol shape that can be inherited
from upstream. Derive discriminators, states, message parts, hook surfaces, and
event unions from the installed SDK's exported types, then lock every member with
an exhaustive
switchorsatisfies Record<UpstreamUnion, ...>. An upstream addition, removal, or rename must fail Stella's typecheck and force an explicit boundary decision. - Derive state types from the SDK or Stella's validated boundary type. Handle them
with an exhaustive
switchplus anevercheck, or asatisfies Record<SdkState, ...>table. - Never use an untyped partial allowlist, truthiness check, or permissive default branch for SDK states. A new SDK member must fail typecheck until its behavior is chosen explicitly.
- Treat values crossing a process, stream, persistence, or browser boundary as untrusted. Reject malformed or unknown wire values deliberately; accept and preserve every documented state, including partial, approval, error, and terminal variants.
- Keep success, recoverable failure, incomplete input, and impossible internal invariants distinct. A valid failed tool call is turn data, not malformed chat.
Preserve lifecycle semantics
- Do not assume one provider iteration or SDK
RUN_STARTED/RUN_FINISHEDpair equals one user-visible assistant message. Tool cycles may contain multiple model runs that contribute to one assistant turn. - Keep a logical assistant message active until the whole turn is complete. Ensure later tool calls, user-input requests, approvals, reasoning, and structured output cannot appear live and then disappear after persistence or refetch.
- End a tool when its promised user-visible result exists. Do not keep an AI tool unresolved across a later, independent UI decision: for example, an editable document draft completes document creation, while saving or exporting that draft is a separate user action. Otherwise the next user message enters a protocol turn whose preceding tool call still has no result.
- Model turn ownership explicitly. Pending user-input and approval tools own the turn until resolved; do not show autonomous follow-ups or drain queued messages while the chat is submitted, streaming, awaiting input, or in error recovery.
- Treat a thread id as a protocol-ownership key. Two independently mounted chat runtimes must not drive the same thread; surfaces that intentionally show one conversation must share one runtime instance. Otherwise allocate a distinct thread so tool continuation, retries, and persistence have one owner.
Show full SKILL.md (364 more words)Show less
Advertise only executable capabilities
- Derive tool schemas and prompt capability instructions from the same executable registry. If a capability cannot succeed in the current context, omit both its tool and its prompt instructions.
- Give public, fixed catalogs exact schemas. Keep dynamic or private catalogs generic only when exposing their values would leak data or create stale schemas.
- Instruct the model to recover from tool errors within the same turn: correct the call, choose an available alternative, or continue without the tool. Escalate to a fatal user-visible error only when the turn itself cannot continue.
- Generate follow-up chips from the user's perspective because they are inserted verbatim as user messages. Suppress them for failed, incomplete, or user-owned turns.
Required tests
- Add a regression test for every fixed AI bug.
- When Stella compensates for an SDK lifecycle or event-shape mismatch, add a canary through the real SDK runtime at that boundary. Assert the user-visible invariant across streamed intermediate events, not dependency-private implementation details, so an upgrade fails where the assumption changes.
- Iterate every SDK union member in state-policy tests. Make the test input type depend on the SDK union so new members cannot be omitted silently.
- For stream changes, test a multi-run tool cycle and assert the browser-visible message parts equal the persisted parts and the reloaded parts.
- Cover success, recoverable tool error, incomplete/streaming structured output, approval, pending user input, cancellation, and malformed wire input where the touched boundary supports them.
- Assert unavailable capabilities are absent from both the final provider tool request and the assembled system prompt.
- For artifact or file surfaces that spawn chat, assert the new surface has empty history, a thread id distinct from its origin conversation, and a stable id across streaming updates and rerenders.
Review checklist
Before finishing, verify all of these:
- Every touched SDK union is exhaustive at compile time.
- Every external boundary validates runtime data without dropping valid states.
- Live, persisted, and reloaded message parts have the same meaning.
- Advertised tools are executable in the current context.
- Recoverable failures let the model continue without UX drama.
- Pending human interactions retain turn ownership.
- Focused tests reproduce lifecycle ordering, not only helper behavior.
© stella, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 1 other file in .agents/skills/conventions-ai of stella/stella.
- SKILL.md
- agents/openai.yaml
Open the folder on GitHubat commit b225fd8
Compare with similar skills
Conventions AI next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Conventions AI this skillstella/stella | 258 | — | ~1.6k | Automated safety check: Pass | Apache-2.0 | |
| Planning With Filesjarrodwatts/claude-code-config | 1.1k | 5 repos | ~967 | Automated safety check: Pass | None | |
| Tool Use Data Synthesissunny-glow/Auto-BenchMax | 1.3k | — | ~3.3k | Automated safety check: Pass | None | |
| Agent Harness ConstructionKartikLabhshetwar/mind-mentor | 148 | 6 repos | ~500 | Automated safety check: Pass | Apache-2.0 | |
| Prompt Engineering Patternswshobson/agents | 40k | — | ~1.3k | Automated safety check: Pass | MIT | |
| Model Benchmarkstheopenco/llmgateway | 1.7k | — | ~1.1k | Automated safety check: Notes | Custom licence |
Similar skills
Tool Use Data Synthesis
sunny-glow/Auto-BenchMax
Synthesize training data for ANY tool-use / agentic benchmark, in ANY repo.
Agent Harness Construction
KartikLabhshetwar/mind-mentor
Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates.
Prompt Engineering Patterns
wshobson/agents
Reference for designing and tuning production LLM prompts: few-shot examples, chain-of-thought, structured outputs, templates and system prompts.
Model Benchmarks
theopenco/llmgateway
Run and report repository model or provider-mapping benchmarks.
Agent Prompt Quality Bar
mastra-ai/mastra
Universal quality bar and final audit rubric for any agent system prompt.
More from stella/stella
All 24 skills in this repoPlan
stella/stella
Create a concise, evidence-backed implementation plan in the repository planning area when the user explicitly asks for a plan.
Answer From Sources
stella/stella
Answers data-protection (GDPR) questions grounded in the regulation and supervisory guidance, with a citation for every claim.
Check Against Rules
stella/stella
Reviews a non-disclosure agreement against the firm's NDA checklist and reports findings with citations.
Intake To Draft
stella/stella
Collects the facts of an unpaid invoice, then drafts a payment demand letter.
Conventions Perf
stella/stella
Apply when a performance-guard check (network baseline, bundle baseline, DB query count, loader-prefetch lint, RC bailouts) fails or when touching a hot route/endpoint.
Conventions Use Effect
stella/stella
Apply when writing or reviewing React effects in apps/web. An agent skill from stella/stella.
Related
Categories
Questions about Conventions AI
What does Conventions AI do?
Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups. Conventions AI is an agent skill from stella/stella. Apply when building or reviewing Stella AI chat, model/provider adapters, tools, prompts, streaming, message persistence, approvals, structured output, or AI-generated follow-ups.
When should I use Conventions AI?
Conventions AI fits situations like: tasks that involve Structured output and tool calling.
How do I install Conventions AI in Claude Code?
Run `npx skills add stella/stella --skill conventions-ai -a claude-code`. Or copy the skill folder (.agents/skills/conventions-ai in stella/stella) into .claude/skills/conventions-ai in your project. Claude Code loads it when a task matches its description.
How do I install Conventions AI in Codex?
Run `npx skills add stella/stella --skill conventions-ai -a codex`. Or copy the skill folder (.agents/skills/conventions-ai in stella/stella) into .agents/skills/conventions-ai in your project. Codex loads it when a task matches its description.
Can I use Conventions AI in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add stella/stella --skill conventions-ai -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/conventions-ai, .gemini/skills/conventions-ai, .github/skills/conventions-ai and .opencode/skills/conventions-ai in your project.
What does Conventions AI need to run?
SKILL.md names no scripts, command-line tools or credentials: Conventions AI is instructions for the agent only.
Does Conventions AI access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Conventions AI safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Conventions AI use?
Conventions AI is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Conventions AI use?
About 1.6k tokens (SKILL.md is roughly 6.2k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Conventions AI?
Skills that share tags, products or a category with Conventions AI: Planning With Files (jarrodwatts/claude-code-config, 1.1k stars), Tool Use Data Synthesis (sunny-glow/Auto-BenchMax, 1.3k stars), Agent Harness Construction (KartikLabhshetwar/mind-mentor, 148 stars) and Prompt Engineering Patterns (wshobson/agents, 40k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Conventions AI?
stella (a GitHub organization) maintains it in stella/stella, which has 258 GitHub stars. The repository holds 24 skills in this directory. The repository was last updated on October 9, 2026.
Source: stella/stella on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.