xurl
Use xurl to access AI agent conversations via agents:// URIs. Invoke when the user: gives an agents:// URI, a provider shorthand like provider/..., or a bare thread/session ID; mentions conversations, threads, or sessions from any AI coding agent; wants to search, read, summarize, compare, or continue agent threads; asks what they worked on, what an agent said, or references past agent interactions; wants to delegate work to or start a conversation in another agent.
What this skill does
## When to Use Trigger this skill when the user's intent involves **any** AI agent conversation — past, present, or to be created. Examples of natural language triggers: - Gives an `agents://` URI, provider shorthand (`provider/...`), or a bare thread/session ID - "Read/show/open this thread/session/conversation ..." - "What did I discuss with [agent] about ...?" - "Summarize my last [agent] session" - "What was I working on in [agent]?" - "Search my agent history for ..." - "Find the conversation where I fixed the auth bug" - "Continue my [agent] conversation about ..." - "Send this to [agent] for review" / "Ask [agent] to ..." - "What subagents were spawned in that thread?" - "Compare what different agents suggested" - "Check if I've discussed X before across agents" `[agent]` can be any AI coding agent name (e.g. codex, claude, copilot, cursor, etc.). xurl supports a growing list of providers — just try it. If a provider is not yet supported, xurl will return a clear error. ## When NOT to Use - General questions about AI agents that don't involve their conversation data - Tasks fully within the current agent session with no cross-agent context needed - Questions about agent capabilities rather than their conversation history ## URI Assembly Guide You are responsible for constructing the correct `xurl` command from the user's input. The user will rarely give a complete `agents://` URI — you must assemble it. ### Decision Flow ``` User input → What do I have? → What do I need? → Construct URI → Run xurl ``` **Step 1: Identify the operation** | User wants to... | Operation | Required info | |---|---|---| | Find/list/search conversations | Query | provider OR path; optional keyword | | Read/show/summarize a conversation | Read | provider + conversation ID | | Inspect metadata or list children | Discover | provider + conversation ID | | Start a new conversation | Write (create) | provider; optional role | | Continue an existing conversation | Write (append) | provider + conversation ID | **Step 2: Resolve missing information** | You have | You're missing | Action | |---|---|---| | Nothing | Provider + ID | Query by path: `xurl 'agents://.?q=<keyword>'` to search current project across all providers | | Provider only | Conversation ID | Query the provider: `xurl <provider>` or `xurl '<provider>?q=<keyword>'`, then pick from results | | Bare thread ID only | Provider | Query by path with the ID as keyword: `xurl 'agents://.?q=<id_fragment>'`; or ask the user which provider | | Provider + keyword | Conversation ID | Search: `xurl '<provider>?q=<keyword>'`, pick matching ID from results | | Provider + ID | Nothing | Ready — construct URI directly | **Step 3: Construct and run** Assemble the URI: `agents://<provider>/<conversation_id>` (or shorthand `<provider>/<conversation_id>`). ### Examples User says: _"Read thread 019c871c-b1f9-7f60-9c4f-87ed09f13592"_ → You have a bare ID but no provider. Search: `xurl 'agents://.?q=019c871c'`, identify the provider from results, then: `xurl <provider>/019c871c-b1f9-7f60-9c4f-87ed09f13592` User says: _"What did I discuss about refactoring in codex?"_ → You have provider (codex) + keyword (refactoring). Search: `xurl 'codex?q=refactoring'`, pick the best match, then read: `xurl codex/<id>` User says: _"Summarize my last copilot session"_ → You have provider (copilot). List recent: `xurl copilot`, take the first result, then read: `xurl copilot/<id>` User says: _"Have codex review this patch"_ → Write operation with provider (codex). Create: `xurl codex -d "Review this patch"` (or with role: `xurl codex/reviewer -d "Review this patch"`) User says: _"Check if I've discussed the migration across any agent"_ → Cross-agent search. Query: `xurl 'agents://.?q=migration'` ## Prerequisites Verify xurl is installed before running any command: ```bash xurl --version ``` If not found, install via the method matching the user's environment: ```bash brew tap xuanwo/tap && brew install xurl # Homebrew cargo install xurl-cli # Cargo / Rust uv tool install xuanwo-xurl # Python / uv npm install -g @xuanwo/xurl # npm / Node ``` ## Workflows ### 1. Query — Find Conversations List recent threads from a provider: ```bash xurl codex ``` Search by keyword with optional limit (default 10): ```bash xurl 'agents://codex?q=refactor&limit=5' ``` Search by project directory (across providers): ```bash xurl 'agents://.?q=migration' # current directory xurl 'agents:///Users/alice/work/repo?limit=5' # absolute path xurl 'agents://~/work/repo?providers=codex,claude' # filter providers ``` Query by role: ```bash xurl codex/reviewer ``` Query results include reduced thread metadata (e.g. `payload.git.branch`, `cwd`) for quick inspection. ### 2. Read — Display a Conversation ```bash xurl codex/<conversation_id> ``` Output is Markdown: YAML frontmatter (metadata) followed by numbered timeline sections (User/Assistant message pairs). Save to file: ```bash xurl -o /tmp/conversation.md codex/<conversation_id> ``` ### 3. Discover — Inspect Metadata and Children ```bash xurl -I codex/<conversation_id> ``` Returns frontmatter with flattened metadata and discovery links (`subagents`, `entries`). Use returned URIs for drill-down: ```bash xurl codex/<main_id>/<child_id> ``` ### 4. Write — Start or Continue Conversations Create: ```bash xurl codex -d "Start a new conversation" xurl codex/reviewer -d "Review this patch" ``` Append: ```bash xurl codex/<conversation_id> -d "Continue with the next step" ``` With provider CLI parameters: ```bash xurl "agents://codex?cd=%2FUsers%2Falice%2Frepo&model=gpt-5" -d "Review this" ``` Payload from file or stdin: ```bash xurl codex -d @prompt.txt cat prompt.md | xurl claude -d @- ``` ## Multi-Step Patterns **Find and read a conversation:** `xurl 'codex?q=<keyword>'` → pick ID → `xurl codex/<id>` **Explore subagents:** `xurl -I codex/<id>` → find child links → `xurl codex/<id>/<child_id>` **Cross-agent project search:** `xurl 'agents://.?q=<keyword>'` → read from whichever provider matches **Resolve a bare thread ID:** `xurl 'agents://.?q=<id_fragment>'` → identify provider → `xurl <provider>/<id>` ## Command Reference ``` xurl [OPTIONS] <URI> ``` | Flag | Purpose | |------|---------| | `-I, --head` | Frontmatter/discovery only (cannot combine with `-d`) | | `-d, --data <DATA>` | Write payload; repeatable; `-d "text"`, `-d @file`, `-d @-` | | `-o, --output <PATH>` | Write output to file | Multiple `-d` values are newline-joined. Path-scoped URIs are read/query only (not valid write targets). ## URI Quick Reference ``` [agents://]<provider>[/<token>[/<child_id>]][?<query>] ``` | Pattern | Operation | Example | |---------|-----------|---------| | `<provider>` | Query recent | `xurl codex` | | `<provider>?q=...` | Keyword search | `xurl 'codex?q=bug'` | | `<provider>/<id>` | Read conversation | `xurl codex/<uuid>` | | `<provider>/<role>` | Role-scoped query | `xurl codex/reviewer` | | `<provider>/<id>/<child>` | Read subagent | `xurl codex/<uuid>/<child>` | | `<provider>` + `-d` | Create conversation | `xurl codex -d "..."` | | `<provider>/<role>` + `-d` | Create with role | `xurl codex/reviewer -d "..."` | | `<provider>/<id>` + `-d` | Append to conversation | `xurl codex/<id> -d "..."` | | `/abs/path` or `.` or `~` | Path-scoped query | `xurl 'agents://.?q=test'` | Token resolution: `<token>` is parsed as session ID first; if that fails, treated as role name. Query parameters: `q=<keyword>`, `limit=<n>` (default 10), `providers=<name,...>` (path-scoped only). In write mode, extra params are forwarded as `--<key> <value>` to the provider CLI. ## Failure Handling xurl returns clear error messages. Act on them directly: | Error | Recovery | |-------|----------| | `command not found: xurl` | Install using Prerequisites section | | `command not found: <agent>` | The provider CLI is not
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.