User Security Common Manual Tasks logo

User Security Common Manual Tasks

Organization
oracle
User Security Common Manual Tasks

Classifies common Oracle Fusion Security Console and Functional Setup Manager tasks by execution type, provides manual completion guidance, and routes approved ESS tasks, including LDAP synchronization processes, to the shared ESS Job Submitter.

Overview

Publisheroracle
RepositoryFusion-Setup-Configuration-Registry
Skill nameUser Security Common Manual Tasks
Stars
112
Forks
26
Bundled files
8
LicenseUPL-1.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 8 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by oracle on GitHub. Read the source before you install it.

Installation

Install the User Security Common Manual Tasks AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/oracle/Fusion-Setup-Configuration-Registry/tree/main/pillars/common/security-console/user-security-common
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/oracle/Fusion-Setup-Configuration-Registry.git /tmp/Fusion-Setup-Configuration-Registry
mkdir -p .claude/skills
cp -r /tmp/Fusion-Setup-Configuration-Registry/pillars/common/security-console/user-security-common .claude/skills/oracle-user-security-common-manual-tasks
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable User Security Common Manual Tasks in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads User Security Common Manual Tasks on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so User Security Common Manual Tasks is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

User Security Common Manual Tasks

Purpose

Classify the currently loaded common user-security FSM tasks, direct the implementor to complete Manual tasks in the Oracle Fusion application, and route approved ESS tasks to the shared ESS job capability.

Cross-Pillar Catalog Scope

This is a multi-task Common skill used by all pillars. It intentionally has no singular task_code. The manifest lists its six catalog-backed IDs in required_setup_tasks; the linked flat task metadata below pairs each ID with its exact name and delegates to this skill. The existing runbook resolver can discover these files without changes outside Security Console. The Current Task Classification table defines their manual and ESS execution boundaries. The two named LDAP processes have no verified FSM code in the effective catalog; their ESS routes remain available by explicit process request.

Setup Tasks

Dedicated setup-object changes

For create, update, delete or redesign requests concerning roles, duties, data-security policies, HCM provisioning rules or LDAP account options, route through common.security-console to that object's focused FSM skill before applying the Manual/ESS classification below. Do not return an unconfigured-execution blocker for a change owned by a configured focused skill. Use FSM input preparation when the user cannot identify, extract or populate the input. This handoff does not authorize an unsupported delete action or bypass Data Loader approval.

Canonical ESS Delegation Contract

Validate schemas/ess-delegation-contract.schema.json version 1.0 before processing an ESS request. Pillars send classification_request with task identity, environment, overrides, and polling options; they cannot supply a job contract or approval. This skill owns classification, configured job defaults, payload construction, preview, and the approval gate. It emits an approved_submission only after explicit approval, with the normalized job_contract, exact request_payload, and approval payload hash. Invoke core.fusion-api.ess-job-submitter only with that approved envelope.

Execution Boundary

For a task classified as Manual, state that it must be completed in the Oracle Fusion application. Do not call Fusion REST APIs, use FBDI or HDL, submit an ESS job, run CSV import or export, or change Fusion configuration.

For a task classified as ESS, use core.fusion-api.ess-job-submitter only after the user provides the target environment, the exact Oracle jobDefinitionId and request parameters or an applicable configured default, and explicit approval to submit. Do not derive an ESS job definition or parameters from an FSM task code.

Do not infer an ESS or other execution type from catalog CSV or export/import flags. Those flags describe catalog support, not an approved FSCR execution classification.

Current Task Classification

Execution typeTaskTask codeCatalog CSV/export-import supportInstruction
ManualManage Configuration Set Migration Target Security PolicyORA_FND_SET_CSM_TARGET_SECURITYNo / NoComplete the task in the Oracle Fusion application.
ManualManage Help Security GroupsATK_MANAGE_HELP_SECURITY_GROUPSYes / YesComplete the task in the Oracle Fusion application; this skill does not authorize CSV processing.
ManualManage Applications Security PreferencesORA_ASE_MANAGE_SECURITY_PREFERENCESYes / YesComplete the task in the Oracle Fusion application; this skill does not authorize CSV processing.
ESSRun User and Roles Synchronization ProcessPER_USER_AND_ROLES_SYNCHRONIZE_DETAILSNo / NoUse the configured default contract below unless the user supplies an override; validate parameters and obtain approval before submission.
ESSImport Users and Roles into Application SecurityORA_ASE_IMPORT_USERS_AND_ROLESNo / NoUse the configured default contract below unless the user supplies an override; validate parameters and obtain approval before submission.
ESSImport User Login HistoryORA_ASE_IMPORT_USER_LOGIN_HISTORYNo / NoUse the configured default contract below unless the user supplies an override; validate parameters and obtain approval before submission.
ESSRetrieve Latest LDAP ChangesTBDNo / NoUse the configured default contract below unless the user supplies an override; validate parameters and obtain approval before submission.
ESSSend Pending LDAP RequestsTBDNo / NoUse the configured default contract below unless the user supplies an override; confirm no active instance, then validate the target environment and obtain approval before submission.
OtherNone currently loadedTBDTBDDo not infer or execute another method. Add a task only after its approved execution boundary is confirmed.

Default ESS Contract: User and Role LDAP Synchronization

For Run User and Roles Synchronization Process and Retrieve Latest LDAP Changes, use the following default Scheduler REST metadata unless the user explicitly provides a replacement value:

FieldDefault value
jobDefinitionIdJobDefinition://oracle/apps/ess/hcm/users/SyncRolesJob
applicationEarHcmEss

The default metadata does not supply request parameters, scheduling values, a target environment, required privileges, or approval. Confirm those inputs before submission. If the user overrides either default value, treat the replacement as a new job contract: validate it against the target environment and present the resulting exact payload for explicit approval immediately before submission.

Default ESS Contract: Import User Login History

For Import User Login History, use the following default Scheduler REST metadata unless the user explicitly provides a replacement value:

FieldDefault value
jobDefinitionIdJobDefinition://oracle/apps/ess/hcm/users/AseInactiveUsersDataLoadJob
applicationEarHcmEss
requestParameters[]

The default metadata does not supply scheduling values, a target environment, required privileges, or approval. Use the empty parameter array unless the user explicitly supplies replacement request parameters. Validate any replacement job definition, application, or parameters against the target environment and present the resulting exact payload for explicit approval immediately before submission.

Default ESS Contract: Import Users and Roles into Application Security

For Import Users and Roles into Application Security, use the following default Scheduler REST metadata unless the user explicitly provides a replacement value:

FieldDefault value
jobDefinitionIdJobDefinition://oracle/apps/ess/hcm/applicationsSecurity/AseImportUsersAndRolesJob
applicationEarHcmEss

The default metadata does not supply request parameters, scheduling values, a target environment, required privileges, or approval. Confirm those inputs before submission. If the user overrides either default value, treat the replacement as a new job contract: validate it against the target environment and present the resulting exact payload for explicit approval immediately before submission.

Default ESS Contract: Send Pending LDAP Requests

For Send Pending LDAP Requests, use the following default Scheduler REST metadata unless the user explicitly provides a replacement value:

FieldDefault value
jobDefinitionIdJobDefinition://oracle/apps/ess/hcm/users/ProcessLdapRequests
applicationEarHcmEss
requestParameters[]

The default metadata does not supply scheduling values, a target environment, required privileges, or approval. Use the empty parameter array unless the user explicitly supplies replacement request parameters. Validate any replacement job definition, application, or parameters against the target environment. Before submission, confirm that no Send Pending LDAP Requests instance is active and present the resulting exact payload for explicit approval immediately before submission.

Scheduled ESS Request Example

For a job whose confirmed contract supports scheduling, use a schedule object only when the user explicitly approves the recurring schedule. Resolve the placeholder values when preparing the exact payload; do not submit the placeholders literally.

json
"schedule": {
  "timezone": "{current_iana_timezone}",
  "recurrences": [
    {
      "icalString": "FREQ=DAILY;INTERVAL=1;",
      "startDate": "{current_timestamp_in_current_timezone}",
      "endDate": "{current_timestamp_in_current_timezone_plus_10_calendar_years}"
    }
  ]
}

Validate the resolved IANA time zone, recurrence, start and end dates, and the job's scheduling support before submission. Do not include both schedule and scheduleId in the same request. This is an input example, not a default schedule; in particular, do not create a recurring Send Pending LDAP Requests schedule unless separately approved.

ESS Submitter Result Handoff

For every ESS invocation, collect the complete structured outputs returned by core.fusion-api.ess-job-submitter:

  • execution_result, including requestId, submission state, and status;
  • diagnostics, including the polling window, poll interval, and timeout result; and
  • audit_summary, including the target environment and any returned external identifiers.

For a live submission, invoke the core skill with the approved polling timeout and interval and a unique --output path in the active project workspace. Do not use --submit-only for an execution initiated by this skill. Wait for the core invocation to finish, then collect and parse the structured output at that path before deciding whether an ESS requestId or returned processId exists. A launched command or quiet console output is not evidence that the job was submitted.

If the invoking client reports completion but the expected output artifact is not immediately available, perform a bounded local collection reconciliation: check for and parse the artifact for up to 30 seconds after client completion. This is a filesystem-read wait only; it must not resubmit the request, reauthenticate, make another ESS status call, or add another polling loop. Capture the core client's exit code and parse any artifact it wrote before classifying the result, because the core submitter writes its structured result before returning its timeout or terminal-failure exit code. Only after this collection window expires with no artifact and no captured ESS identifier may the outcome be treated as indeterminate.

The core submitter writes a redacted failure artifact whenever its --output path is supplied, including failures before a request ID is available. Parse that artifact first and persist its failure class and message. Treat a missing artifact after the reconciliation window as artifact_persistence or client handoff failure; it is not evidence that the ESS request was not submitted and must never trigger an automatic resubmission.

Unidentified Submission Recovery

When a completed submission handoff has no request ID or process ID, do not retry or automatically query Fusion. Return this user-facing response after persisting the redacted failure artifact: “I couldn't find a request ID. I can check whether a matching job ran in the target environment during the last few minutes. Do you want me to check?”

If the user explicitly asks to check, use the core submitter's read-only --find-recent-job-definition and --since-minutes route with the exact approved job definition, target environment, and user-requested window (for example, two or three minutes). Persist every matching request ID and status. If exactly one candidate matches, return its ID and offer read-only status tracking. If zero candidates, multiple candidates, inaccessible history, or a lookup failure occurs, do not submit again. Tell the user: “I couldn't find an unambiguous matching job. Please open Scheduled Processes in the target Fusion application and check for this job before trying again.”

Treat execution_result.requestId as the Oracle Scheduler request/process identifier. If Fusion returns a separate processId in a full status payload, preserve it as an additional external identifier; do not invent one when it is absent. Persist the returned result in the project execution audit, redacting credentials, tokens, secrets, and sensitive parameter values. Include at least the request ID, job definition, application, target environment, poll settings, last or terminal status, process phase, cause, elapsed time, ECID when present, and references to available diagnostics.

Use core.fusion-api.ess-job-submitter as the sole authority for submitting, tracking, OAuth refresh-once behavior for HTTP 401 or 403, polling, timeout, and terminal-status handling. Do not add a second polling loop, refresh, or live retry in this skill. Apply the core skill's result rules as follows:

  • On its successful result, persist the returned request ID and structured outputs.
  • On its exit code 3, classify the result as process_timeout, retain the returned request ID and last status, and do not retry.
  • On its exit code 4, classify the result as process_failed, retain the returned request ID and terminal status, and do not retry.
  • If the completed core invocation returns no ESS requestId or processId, classify the outcome as indeterminate. Preserve the structured result and available diagnostics, and do not retry the live request until its status is confirmed and fresh approval is received.

ESS Input Confirmation Gate

Before every ESS submission, resolve and display the complete submission input to the user once. The display must include the target environment, operation, full payload (including defaults and overrides), request parameters, schedule or on-demand mode, polling timeout and interval, and any relevant preflight condition such as the no-active-instance check for Send Pending LDAP Requests.

Ask the user to confirm that displayed submission. Run the job only after that confirmation is received. A confirmation of the displayed submission is the single user-confirmation gate for those inputs; do not ask the user to confirm the same unchanged inputs again. If an input, target, payload, schedule, or polling setting changes after confirmation, display the revised submission and obtain a new confirmation before running it.

Required Response: Manage Configuration Set Migration Target Security Policy

When the user asks to set Manage Configuration Set Migration Target Security Policy, respond with the following message, adapting only the project and environment names when they are known:

Manage Configuration Set Migration Target Security Policy (ORA_FND_SET_CSM_TARGET_SECURITY) is a Manual FSM task. Complete it in the Oracle Fusion application; this FSCR skill must not use REST APIs, CSV import/export, FBDI, HDL, or ESS to set the policy.

Before making the change, confirm the target Fusion environment, the configuration set and migration target in scope, and the approved security policy values from the implementation security design. Do not infer policy values or widen user access.

In the Oracle Fusion application, open the relevant Setup and Maintenance task from the implementation project, review the current target security policy, apply the approved values, and save the change. Follow the project's approval process before applying a policy that changes access to migrated setup data.

After completion, provide sanitized evidence: the environment, configuration set or migration-target identifier, completion status, completion time, and reviewer/approver. Do not include credentials, tokens, customer data, or sensitive policy values in the evidence.

TBD: Exact page fields, allowed policy values, and validation criteria must come from approved Oracle product guidance and the project security design.

Required Response: ESS Tasks

When the user asks to run one of the ESS tasks in the Current Task Classification table, respond with the following message, adapting only the task name and code:

{Task Name} ({Task Code}) is an ESS task. FSCR will use core.fusion-api.ess-job-submitter to submit and monitor the process only after you provide the target Fusion environment, request parameters, and explicit approval to submit. For Run User and Roles Synchronization Process and Retrieve Latest LDAP Changes, FSCR defaults to JobDefinition://oracle/apps/ess/hcm/users/SyncRolesJob and EarHcmEss unless you explicitly override either value. For Import Users and Roles into Application Security, FSCR defaults to JobDefinition://oracle/apps/ess/hcm/applicationsSecurity/AseImportUsersAndRolesJob and EarHcmEss unless you explicitly override either value. For Import User Login History, FSCR defaults to JobDefinition://oracle/apps/ess/hcm/users/AseInactiveUsersDataLoadJob, EarHcmEss, and requestParameters: [] unless you explicitly override a value.

The FSM task code is not an ESS job definition and must not be used as one. TBD: scheduling requirements and completion criteria must be confirmed from Oracle product guidance or target-environment metadata before submission. Any job definition, application, or request-parameter override must also be confirmed before submission.

After approval, FSCR will submit the request, record the returned request ID, monitor it within the approved timeout, and persist a redacted execution audit. It will not retry a failed or timed-out ESS submission without a new approval or an idempotency-safe recovery path.

Required Response: Send Pending LDAP Requests

When the user asks to run Send Pending LDAP Requests, respond with the following message, adapting only known project and environment details:

Send Pending LDAP Requests is an ESS process. It sends pending bulk and future-dated user-account and role-provisioning or deprovisioning requests to LDAP. FSCR will use core.fusion-api.ess-job-submitter only after you provide the target Fusion environment, required privileges, and explicit approval immediately before submission. FSCR defaults to JobDefinition://oracle/apps/ess/hcm/users/ProcessLdapRequests, EarHcmEss, and requestParameters: [] unless you explicitly override a value.

The display name is not an ESS job definition ID. TBD: Confirm any parameter override, batch-size rules, scheduling requirements, and completion criteria from approved Oracle product guidance or target-environment metadata before submission. Do not infer them from an FSM task code.

Before submission, confirm that no existing instance of Send Pending LDAP Requests is running. Only one instance may run at a time. Submit an on-demand request unless the user explicitly approves a schedule; do not create a recurring schedule by default.

After approval, FSCR will submit and monitor the request, record the request ID, target environment, redacted parameters, terminal status or timeout, elapsed time, and diagnostic references in the project workspace. It will not retry a failed or timed-out request without fresh explicit approval.

Workflow

  1. First route dedicated setup-object changes as described above. For the remaining common controls, match the task to the Current Task Classification table.
  2. For Manage Configuration Set Migration Target Security Policy, use the Required Response section exactly.
  3. For either of the other two Manual tasks, respond: This task is Manual. Complete it in the Oracle Fusion application.
  4. For an ESS task, use the Required Response: ESS Tasks section. For Run User and Roles Synchronization Process, Retrieve Latest LDAP Changes, Import Users and Roles into Application Security, Import User Login History, or Send Pending LDAP Requests, apply the applicable Default ESS Contract unless the user provides an override. For Send Pending LDAP Requests, use the Required Response: Send Pending LDAP Requests section. Validate the task, target environment, jobDefinitionId, application, request parameters, privileges, timeout, and polling interval. Display the complete resolved submission inputs once and obtain the user's confirmation of that display before invoking core.fusion-api.ess-job-submitter. For a scheduled request, also validate the schedule's resolved time zone, recurrence, start and end dates, and explicit schedule approval; do not use both schedule and scheduleId.
  5. Invoke core.fusion-api.ess-job-submitter using its approved polling options and a unique project-workspace --output path. After the client returns, collect its execution_result, diagnostics, and audit_summary from that artifact. If the artifact is not immediately available, use the bounded local collection reconciliation in ESS Submitter Result Handoff before classifying the result. Require its returned ESS request ID or process ID before treating the submission as identified, and apply its polling, timeout, and terminal failure rules without adding a second poll or retry. Record the request ID or returned process ID, target environment, job definition, application, polling settings, terminal or last status, and redacted diagnostics in the project workspace when project context is supplied. Treat the invocation as indeterminate only after the collection window expires with no structured result or ESS identifier; do not retry without a confirmed status and fresh approval.
  6. Before submitting Send Pending LDAP Requests, confirm that no existing instance is running. Allow an on-demand request or an explicitly approved schedule only; do not create a recurring schedule by default.
  7. Identify the task name and code, and request completion evidence appropriate to the implementation process, such as a reviewer-confirmed completion status or sanitized screenshot.
  8. If the task is listed as Other, state its configured boundary. Do not invoke a job, API, loader, or another skill unless the user explicitly asks for an approved capability.
  9. If the task is not listed, report TBD: execution type is not configured for this skill. Do not infer a manual, ESS, or automated route.

Completion Guidance

For the currently loaded Manual tasks, the implementor must use the relevant Oracle Fusion setup or Security Console page and follow approved product and project procedures. This skill provides task classification only; it does not define product-field values, make design decisions, or verify that a manual configuration is correct.

Preflight Checks

Before providing manual-completion guidance, confirm that the requested task has an unambiguous classification in the Current Task Classification table. For an ESS task, before invoking core.fusion-api.ess-job-submitter, confirm:

  • target Fusion environment identity;
  • exact Oracle jobDefinitionId, application, and request parameters; use an applicable Default ESS Contract only for its named task and only until the user explicitly overrides it;
  • required Fusion privileges and the approved timeout and polling window;
  • project workspace destination, when project context is supplied; and
  • one user confirmation of the displayed complete submission input, including the exact target environment, operation, payload, and polling settings.

Fail closed when the task, job contract, target environment, or approval is missing, ambiguous, or mismatched. Do not use an FSM task code or display name as a substitute for an ESS job definition. For Send Pending LDAP Requests, also confirm no instance is already running before submission.

Failure Handling

Classify and report failures as input_validation, catalog_or_metadata_mismatch, authorization_or_privilege, transport_or_environment, fusion_api_error, process_timeout, process_failed, artifact_persistence, or approval_blocker.

  • For an invalid or incomplete ESS contract, environment mismatch, missing privilege, or approval blocker, stop before any side effect and state the required correction.
  • For Fusion API failures, record the endpoint path, response summary, and request ID or ECID when available; redact credentials, tokens, customer data, and sensitive configuration values.
  • For a timeout or failed ESS process, persist the target environment, job definition, request ID, last status, elapsed wait, and available diagnostics in the project workspace when project context is supplied.
  • Do not supplement the core submitter's polling, authentication refresh, or timeout behavior with a separate request loop. Preserve the core result and use its returned request ID to resume tracking only after the required approval gate is satisfied.
  • Do not retry a failed, timed-out, or indeterminate ESS submission without a confirmed request status and fresh explicit approval, unless an idempotency-safe recovery path is confirmed.
  • If audit or diagnostic persistence fails, return an artifact_persistence failure and identify the intended artifact location without exposing sensitive values.

Safety Rules

  • Do not expose credentials, tokens, secrets, customer data, or sensitive security configuration in completion evidence.
  • For a common control classified Manual, hand off to the implementor. Dedicated FSM setup changes use their focused skill and Data Loader approval.
  • Do not submit Send Pending LDAP Requests while another instance is running.
  • Do not create a recurring Send Pending LDAP Requests schedule without explicit user approval.
  • Do not retry a failed or timed-out Send Pending LDAP Requests request without fresh explicit approval.
  • Mark unavailable task types, product behavior, and source guidance as TBD rather than inventing a route.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the User Security Common Manual Tasks AI skill do?

Classifies common Oracle Fusion Security Console and Functional Setup Manager tasks by execution type, provides manual completion guidance, and routes approved ESS tasks, including LDAP synchronization processes, to the shared ESS Job Submitter.

Why use User Security Common Manual Tasks on TypingMind?

Because you install it once and use it with any model. User Security Common Manual Tasks is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install User Security Common Manual Tasks in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/oracle/Fusion-Setup-Configuration-Registry/tree/main/pillars/common/security-console/user-security-common. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use User Security Common Manual Tasks?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with User Security Common Manual Tasks?

As many as you like. As long as a model supports skills, you can use User Security Common Manual Tasks with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the User Security Common Manual Tasks AI skill free?

Yes. It is published on GitHub by oracle under the UPL-1.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇