ExtendRoom 12 of 13
Every tool declares its risk before it runs.
The catalog covers 63 skills across developer workflows, AI agents, office productivity, media processing, and more. Install more from the Marketplace or let the agent author its own.
Each of the 132 tools cataloged here carries a risk level in its manifest, and the approval gate reads it on every call.
A recreation of the File System skill under an agent's tools in the Sophon Dashboard: 9 tools, each with its declared risk. The agent calls fs.read (None), fs.copy (Low), fs.write (Medium), fs.delete (High). At the default Medium threshold, None and Low run, and Medium and High stop and ask you first.
04 / 04 · fs.delete · High · asks you
Recreation of an agent's tools in the Dashboard. Verdicts at the default Medium threshold.
AWatch a skill go from package to run
One skill, from package to finished run.
A skill folder with manifest.json, SKILL.md and main.py. The manifest declares the tool my_tool with risk level low and a network allowlist of api.github.com.
Step 1 of 5: A skill is a package
01 / 05 · A skill is a package
A
manifest.jsonfor the machine: tool names, parameters as JSON Schema, risk levels, required credentials, sandbox config. ASKILL.mdfor the agent: what the skill does and when to use it. Then the code, in Python or C#.The risk level is written down before the tool ever runs.
BMechanism
Five risk levels, written into the manifest.
- Risk: None
Read-only, deterministic, cannot affect anything.
Runs.
44 cataloged tools · memory.search, fs.read
- Risk: Low
Reads with side effects, or minor writes.
Runs.
38 cataloged tools · web.search, fs.copy
- Risk: Medium
Writes data, sends messages you could undo.
Asks you at the default threshold.
35 cataloged tools · fs.write, message.send
- Risk: High
External, costly, or hard to reverse.
Asks you at the default threshold.
14 cataloged tools · fs.delete, email.send
- Risk: Critical
Destructive or privileged.
Always asks, whatever the threshold or any override says.
1 cataloged tool · agent.delete
Overrides, per-call assessment and the Critical floor are applied in that order. How the gate resolves a call
CCatalog
Every skill, every tool, every level.
Category
63 skills · 132 tools
github.reposRisk: NoneList user repositories
github.issuesRisk: NoneList issues for a repository
github.create_issueRisk: MediumCreate a new issue
jira.searchRisk: NoneSearch Jira issues using JQL
jira.get_issueRisk: NoneGet details of a specific issue
jira.create_issueRisk: MediumCreate a new Jira issue
jira.list_projectsRisk: NoneList all accessible projects
confluence.searchRisk: NoneSearch pages and content using CQL
confluence.get_pageRisk: NoneGet a page by ID with content
confluence.create_pageRisk: MediumCreate a new page
confluence.update_pageRisk: MediumUpdate an existing page
confluence.list_spacesRisk: NoneList all accessible spaces
confluence.add_commentRisk: LowAdd a comment to a page
email.sendRisk: HighSend an email
email.searchRisk: NoneSearch emails by query
google.mail_listRisk: NoneList recent inbox emails
google.mail_readRisk: NoneRead a specific email by ID
google.mail_searchRisk: NoneSearch emails by query
google.mail_sendRisk: HighSend a new email
google.mail_replyRisk: HighReply to an email
google.calendar_listRisk: NoneList upcoming calendar events
google.calendar_createRisk: MediumCreate a calendar event
google.calendar_updateRisk: MediumUpdate a calendar event
google.calendar_deleteRisk: HighDelete a calendar event
google.calendar_searchRisk: NoneSearch calendar events
browser.actRisk: MediumPerform browser actions: navigate, click, type, screenshot, scroll, evaluate JS, extract text, PDF
browser.screenshotRisk: LowTake a screenshot of a web page
web.searchRisk: LowSearch the web for information
web.scrapeRisk: LowScrape pages with CSS/XPath selectors, screenshots, PDF generation, readability extraction
web.feedRisk: LowParse RSS/Atom feeds and sitemaps. Fetch, filter, and merge feeds
http.requestRisk: MediumMake an HTTP request to a URL
code.executeRisk: MediumExecute code in a sandboxed environment (Python, C#)
code.buildRisk: MediumBuild a project (dotnet, npm, pip, cargo, go)
code.testRisk: MediumRun tests (dotnet test, npm test, pytest, cargo test, go test)
code.lintRisk: LowRun linting and static analysis (eslint, pylint, rustfmt, go vet)
code.formatRisk: LowAuto-format code (prettier, black, rustfmt, gofmt)
code.depsRisk: LowAudit, list, or check dependencies for vulnerabilities
code.gitRisk: MediumGit operations: status, diff, log, commit, branch, checkout, merge, stash, clone, tag, blame, cherry-pick
fs.listRisk: NoneList files and directories
fs.readRisk: NoneRead file contents with line offset/limit
fs.writeRisk: MediumWrite content to a file (overwrite, append, create)
fs.copyRisk: LowCopy files or directories
fs.moveRisk: MediumMove or rename files
fs.deleteRisk: HighDelete files or directories
fs.findRisk: NoneFind files matching glob patterns
fs.infoRisk: NoneGet file/directory metadata
fs.create_directoryRisk: LowCreate directories recursively
archiveRisk: LowCreate, extract, list, add, remove, encrypt, convert archives
document.uploadRisk: LowUpload a file to the document library
document.searchRisk: NoneSearch across uploaded documents
document.askRisk: NoneAsk a question about a document
document.summarizeRisk: NoneSummarize a document
document.listRisk: NoneList all uploaded documents
document.readRisk: NoneRead document content
document.infoRisk: NoneGet document metadata
document.updateRisk: LowUpdate a document's content or metadata
document.collection.createRisk: LowCreate a document collection
document.collection.listRisk: NoneList document collections
document.collection.deleteRisk: MediumDelete a document collection
document.deleteRisk: MediumDelete a document
pdfRisk: MediumCreate, merge, split, rotate, watermark, encrypt, decrypt, fill forms, extract tables/images/links, compress
ocrRisk: NoneExtract text, tables, classify documents, parse receipts/invoices, detect handwriting
markdownRisk: LowConvert to HTML/PDF, render tables, generate TOC, extract links and headings
office.excelRisk: MediumCreate, edit cells/ranges, formulas, charts, pivots, filter, sort, merge sheets, import/export CSV
office.wordRisk: MediumCreate, edit, merge, split, convert, find/replace, TOC, headers/footers, styles
office.powerpointRisk: MediumCreate and manipulate PowerPoint presentations
data.csvRisk: LowParse, filter, sort, aggregate, merge, convert, validate, pivot, deduplicate, join CSV data
data.jsonRisk: LowParse, query, transform, validate, merge, diff, flatten, schema generate, patch, filter JSON
data.sqliteRisk: MediumQuery, export, import, create tables, insert, vacuum, describe SQLite databases
data.xmlRisk: LowParse, XPath query, XSLT transform, XSD validate, convert, merge XML
data.yamlRisk: LowParse, convert to/from JSON, validate, merge, format, query YAML
media.imageRisk: LowResize, crop, rotate, flip, watermark, convert, compress, thumbnail, overlay, text, collage, blur, sharpen
media.audioRisk: LowConvert, trim, merge, split, normalize, waveform, volume adjustment
media.videoRisk: LowTranscode, thumbnail, trim, merge, compress, extract audio, GIF, resize
image.generateRisk: MediumGenerate images from text prompts (DALL-E, Fal.ai, Stability AI)
tts.fileRisk: LowGenerate audio from text, batch generate, list available voices
calculator.evaluateRisk: NoneEvaluate math expressions (+, -, *, /, parentheses)
translateRisk: LowTranslate text, detect language, batch translate, translate files
qrcodeRisk: LowGenerate QR codes, read QR codes, generate barcodes, read barcodes
datetime.nowRisk: NoneGet current date and time, optionally in a specific timezone
cryptoRisk: MediumHash, verify, encrypt/decrypt files, sign, verify signatures, generate keys, checksums, passwords
diffRisk: LowText diff, file diff, generate patches, apply patches, three-way merge
templateRisk: LowRender templates, validate syntax, list variables, batch render
contactsRisk: LowCreate, parse, merge, deduplicate, convert, and search vCard files
calendarRisk: LowCreate events, parse ICS, merge, convert timezones, expand recurring, export
email.composeRisk: LowCreate EML files, parse emails, create calendar invites, extract attachments
message.sendRisk: MediumSend a message to a connected channel (Telegram, WhatsApp, Slack, Discord, Email)
channel.listRisk: NoneList all channels with type, name, and connection status
memory.writeRisk: LowWrite a fact or preference to long-term memory
memory.searchRisk: NoneSearch daily logs, long-term memory, and vector store
memory.listRisk: NoneList memory entries for the current agent
memory.graphRisk: NoneExplore the entity graph: overview, neighbors, path, or cluster
memory.diagramRisk: NoneExport the memory graph as a Mermaid diagram
memory.linkRisk: NoneCreate a link between two memory entities
memory.entity.listRisk: NoneList entities in the memory graph
memory.forgetRisk: MediumDelete a specific memory entry by ID
os.executeRisk: HighExecute a shell command on the host OS (bash, cmd, powershell)
os.execute_sandboxedRisk: MediumExecute a command inside a Docker sandbox
os.infoRisk: NoneGet OS version, architecture, hostname, and available shells
processRisk: MediumList, info, kill processes, resource usage, top view
networkRisk: LowPing, DNS lookup, port check, WHOIS, HTTP trace, IP info
clipboardRisk: LowRead, write, and clear the system clipboard
agent.delegateRisk: MediumDelegate a task to a sub-agent that can work autonomously
agent.listRisk: NoneList available agents
agent.getRisk: NoneGet an agent's definition
agent.templatesRisk: NoneList agent templates
agent.updateRisk: HighUpdate an agent's configuration
agent.deleteRisk: CriticalDelete an agent
canvas.renderRisk: LowPush HTML, charts, tables, code, diagrams to the canvas panel
insights.queryRisk: NoneQuery usage analytics and cost tracking data
cron.manageRisk: MediumCreate, list, update, delete, pause, resume scheduled jobs and view history
workflow.createRisk: HighCreate a workflow from a natural language description
workflow.listRisk: NoneList saved workflows
workflow.describeRisk: NoneDescribe a workflow's steps
workflow.editRisk: HighEdit an existing workflow
workflow.triggerRisk: LowTrigger a workflow run
workflow.pauseRisk: LowPause a workflow
workflow.resumeRisk: LowResume a paused workflow
workflow.deleteRisk: HighDelete a workflow
connection.configureRisk: MediumStore service connection credentials
connection.listRisk: NoneList available service connections and their fields
connection.statusRisk: NoneCheck connection status of a service
music.generateRisk: MediumGenerate music or audio from a text prompt (ElevenLabs, Stability, MiniMax)
video.generateRisk: MediumGenerate a video from a text or image prompt
weatherRisk: LowGet current weather and forecast for a location
coding.claude_codeRisk: HighDelegate a coding task to the Claude Code CLI
coding.codexRisk: HighDelegate a coding task to the OpenAI Codex CLI
coding.copilotRisk: HighDelegate a coding task to the GitHub Copilot CLI
remote.controlRisk: MediumControl a paired desktop: screenshot, click, type, launch apps, manage windows
node.present_canvasRisk: MediumOpen a canvas on the paired desktop's screen
DWhere skills come from
Bundled, installed, or written by the agent.
- Bundled
- Ship with every Sophon install. Enable them per agent in Agents › Tools; newly installed skills start disabled on agents that already exist.
- Marketplace
Community-built skills and plugins for Notion, Todoist, Stripe, Salesforce, HubSpot, Spotify, Home Assistant, and more. Every new version waits for a moderator before it is visible. Install from the Dashboard (Skills › Marketplace) or with
sophon skills install notion; an install the agent starts is always gated behind your approval.- Self-authored
- Ask for a skill in plain language and
skill.author(High risk) writes the code, the manifest with its risk level, and the SKILL.md, tests it in a sandbox with sample parameters, and asks for your approval before installing it.
ESkills or plugins
Skills call tools. Plugins extend the core.
Since 2.2, the Marketplace also carries agent packages: personas and setup you can review, install and customize. Skills supply their tools. Explore agent packages.
A skill gives the agent a new tool, sandboxed. A plugin gives Sophon itself a new channel type or document format. Plugins run out of process with full trust, are admin-gated, and are distributed through the same Marketplace.
Limits
What a risk level doesn't promise.
The declared level is a default, not a floor. An admin's per-tool override can raise it or lower it, except on Critical tools, which always ask.
Risk levelsA self-authored skill picks its own risk level. Review the generated manifest before you approve it: that level decides how often you are asked later.
Self-authoringThe vault keeps a credential out of the model's prompt, not out of the skill. A sandboxed skill holds the real credential while it runs, and its network allowlist is what bounds where it can go. A broker that makes the call on its behalf is roadmap.
Credential vault