Overview
A Skill is a reusable capability bundle. Every skill consists of two parts:- A SKILL.md instruction document that tells the agent “when to use it, how to use it, and what steps to follow”;
- The allowed tools declared by the skill, plus optional supporting resource files (scripts, templates, sub-documents, and so on).
.zip / .skill / .tar.gz / .tgz) and uploaded to AI SRE. The archive’s root directory must contain a SKILL.md file. All other resource files can be placed anywhere inside the archive and will be read by the agent on demand during execution.
Once enabled, the skill becomes visible and callable by the agent in the session. The agent can autonomously decide when to invoke a skill, or you can explicitly trigger it in the chat box with /<skill-name> (for example, /skill-creator).
When triggering explicitly, you can append arguments: /<skill-name> arg1 arg2. In the SKILL.md body, $1…$9 reference whitespace-split positional arguments, and $ARGUMENTS references the full raw remainder. AI SRE substitutes the actual values before it sends the turn.
If you only want to mention /skill-name in a message (for example, asking “what does \/skill-name do?”) without triggering it, add a backslash escape at the start of the message. A message starting with \/ has that leading backslash stripped and is sent as plain text instead of being parsed as a command.
Cross-team skill name conflicts and the
@ qualifier. Skill names must be unique within a team, but different teams may each own a skill with the same name. Two teams in the current session’s scope might both have an enabled skill with the same name. A bare /<skill-name> then can’t resolve to either one. AI SRE rejects it and lists the qualified forms. In that case, use /<skill-name>@<team_id> to disambiguate (for example, /deploy@5). The @ character is therefore reserved and cannot be used in skill names.Built-in Skills
The AI SRE runtime bundles these skills; no installation needed:
There are more, including
init, insight, artifact-design, alert-governance, channel-noise and monit-query. monit-query loads only when the account has an enabled Monitors data source. To see which skills a given session has, ask the agent.
channel-noise: console entry only, and it writes settings
It starts only from a channel’s noise-reduction page. Click Let AI tune noise reduction on Channel details → Noise reduction → Alert grouping or Flapping detection: the AI SRE panel opens with the channel attached as a reference and a /channel-noise question pre-filled, and it starts when you send it. Typing /channel-noise in IM, automation rules, or the Open API is rejected with /channel-noise can only be started from its entry in the console. The agent also does not pick it on its own, so asking in plain language to “look at this channel’s alert noise” does not trigger it.
It writes settings. It reads the channel’s own alert history (the last 30 days by default), locates where the noise comes from, estimates several candidate alert grouping and flapping detection settings, and recommends one. Only after you explicitly confirm does it write to the channel and read the result back. It changes only alert grouping and flapping detection; other noise-reduction settings such as incident merging, silence, inhibition, and drop get suggestions only, for you to adjust. On the page itself it fills the recommended values into the form and they take effect when you click Save; see Operating console pages.
Install from marketplace
The skill Marketplace is part of the plugin catalog on Customize → Plugins → Overview, under the Skill category; search by name or description.1
Find the template
On Customize → Plugins → Overview, find the template under the Skill category; each row shows the template name and description.
2
Confirm the install
Click a template row or its +, then confirm to install. Marketplace installs are always shared scope (available to all members of the account). The template is copied into your account as a regular skill, labeled with its source template and version.
3
Manage what is installed
An installed row no longer shows +; it shows Enabled. Managing and updating it happens on the Customize → Plugins → Skill list — see “Management and Inspection” and “Automatic and Manual Updates” below.
New accounts come with a set of official Marketplace templates pre-installed:
browser-automation (a browser automation CLI for operating websites, dashboards, and monitoring UIs), mcp-builder (guides you through building an MCP server), and skill-creator (see “Add in conversation” below). These pre-installed skills behave like manually installed skills. You can enable/disable, uninstall, or update them to the latest version under “Management and Inspection” below.Automatic and manual updates
When a Marketplace template publishes a newer version, the corresponding skill entry displays an Update available badge. Whether the update is applied automatically depends on whether the skill has been modified locally:Clean skill (unmodified)
A skill that has not been locally modified after installation is automatically updated to the latest Marketplace version when the list loads — no manual action required. A notification appears after a successful update.
Modified skill
A skill that has been locally edited or re-uploaded is skipped by auto-update to avoid overwriting your changes. It retains the Update available badge, and you must manually click to update and confirm the overwrite.
Custom Skills
In addition to installing from the Marketplace, you can upload your own skill packages. The toolbar of the skill list page has two entries:- Add from catalog: jumps to the Skill category of Customize → Plugins → Overview (the catalog described under “Install from marketplace”), where you pick and install a curated skill template; it then appears in this list. It needs the permission to upload skills.
- Upload Skill: opens the upload form. Fill in:
At upload time, the system validates the package. The archive must be a valid zip with
SKILL.md in the root directory and parseable frontmatter. name must follow the naming rule: start with a letter, use only letters/digits/-/_, and be 1–64 characters. All declared tools must be valid, meaning the built-in tools and MCP servers exist. Skill names must be unique within a team — a duplicate name within the same team is rejected with a prompt to choose a different name. Different teams may each own a skill with the same name.
The system reads all metadata other than the skill name from the
SKILL.md frontmatter, including description, version, tags, author, and tools. You don’t enter it in the form.Overwrite upload (replace)
By default, uploading an archive whose name matches an existing skill is rejected. To replace a skill with a new version, choose “Replace” in its detail panel. You can also upload a skill with the same name and confirm the “Replace” dialog. The SkillID stays the same, so references such as/<skill-name> keep working; a replace never creates a new skill.
Through the API, upload with replace=true to match and overwrite by the name in SKILL.md; add a skill_id to overwrite by ID instead (useful after a rename). A missing target returns 404.
Add in conversation (skill-creator)
Besides uploading a zip, you can create and refine skills right inside an AI SRE session.skill-creator is a Flashduty Marketplace skill that builds skills. It’s pre-installed in your account by default. Trigger it with /skill-creator in any session, or describe what you want in natural language:
- Create from scratch: say “help me create a skill for troubleshooting X.” After an investigation, you can say “turn that workflow into a skill.” skill-creator clarifies your intent and drafts the
SKILL.md. It can also set up test cases and iterate on them. When you’re satisfied, it saves the result as a skill in your account. - Install an existing one: click Add in conversation at the top right of the skill list page and choose to install an existing skill. The agent checks the plugin catalog first, then searches the web if nothing matches; once you pick one, it completes the install. (The button appears only when the target scope is unambiguous — you have account-level permission, or exactly one of your teams is selected.)
- Rewrite & optimize: click “Edit in Chat” in a skill’s detail panel to have skill-creator rewrite that skill; it can also tune the
descriptionto improve triggering accuracy.
Management and inspection
The skill list shows each skill’s name, scope (shared or team), version and enabled state, with labels for the source template, Update available and Local runner only. Skills that declarevenues: [byoc] are labeled Local runner only: they work only in sessions bound to a local runner, not in cloud sandbox sessions. Marketplace templates carry the same label so you can tell before installing. You can filter by scope (All / Account / Team) and search by name, description or author.
Enable / Disable
Enable / Disable
Toggle using the switch in the list or the detail panel. Only enabled skills are visible to the agent; a disabled skill is invisible to the agent and cannot be invoked.
Edit
Edit
Click the edit button to update the description and owner (scope). Marketplace-installed skills are permanently shared scope, so the scope is not editable in the form. To modify skill content, download the zip, edit it, then re-upload (this creates a new version and marks the skill as modified).
Replace / Re-upload
Replace / Re-upload
In the detail panel, select “Replace” to overwrite the current skill content with a new zip. The SkillID remains unchanged, so any existing references (for example,
/<skill-name> triggers) continue to work.Download
Download
Download the complete zip package (including
SKILL.md and all resource files) for offline editing or backup.Uninstall / Delete
Uninstall / Delete
Removes the skill from the current scope. Any active sessions currently using it will continue to fail until the skill is reinstalled. This action requires confirmation.
Detail panel
Click any row in the list to open the skill’s detail panel:- Left side: the file tree of the skill package (
SKILL.md,README.md, and so on are prioritized as the default preview file). - Right side: content preview of the selected file, with the skill description shown at the top.
- Header area: skill name and short reference (such as
skill-xxxxxx), scope, source template version, Update available marker, author,versionand the full SkillID. - Actions: try in chat (injects
/<skill-name>into a new session), update to latest version, edit in chat, replace, download, uninstall. - Try asking: if the skill declares
examplesin theSKILL.mdfrontmatter, you can expand up to 2 sample questions. Clicking one opens a new chat prefilled with the skill and question, ready for you to review and send.
Scope
Skills share one set of scope and permission rules with the other customize resources; see Scope and permissions. What is specific to Skills:- Marketplace installs are always shared scope. Any member can install, and you can’t pick a team.
- A few older team-scope Marketplace skills can be promoted to account level with Set to Shared in the detail panel. Only the account owner or admins can do this.
- The
/autocomplete lists only shared-scope skills and skills of teams you belong to; typed slash commands follow the same scope. In a team session, calling another team’s skill is rejected with a reason. - Without edit permission, the row is shown as read-only.
SKILL.md format
SKILL.md consists of two parts — YAML frontmatter (metadata) and a body (human-readable instructions for the agent) — separated by ---:
Use
venues for skills that depend on one kind of environment. For example, a skill that needs a local binary, local files, or a private-network service on the host can’t work inside a Cloud Sandbox. With venues: [byoc] declared, the skill neither appears in the agent’s available-skills list nor can be triggered by /<skill-name> in a Cloud Sandbox session. Only sessions bound to a self-hosted Runner can see and call it.- Built-in tools: write the tool name directly. Available built-in tools include
read,write,edit,bash,grep,glob,skill,mcp,todo,time,webfetch,web_search,session_transcript, and others. - MCP tools: write them as
mcp:server/tool(for example,mcp:my-server/query). At upload time, only the existence of the MCP server is validated; the specific tool name is verified when the MCP is loaded during a session.
Related pages
Manage Knowledge
Use DUTY.md and knowledge to provide the agent with team context and troubleshooting experience.
MCP (External Tools)
Connect external tools so skills can invoke them in SKILL.md using
mcp:server/tool.Agent
Extend AI SRE’s collaboration and task-delegation capabilities with Agents.
Console Chat
Explicitly trigger a skill in a session with
/<skill-name>, or let the agent invoke it autonomously.Overview
Learn about AI SRE’s overall capabilities and positioning.