Skip to content

{{PRODUCT_NAME}} Agent Skills for Coding Agents

Agent skills give a coding agent a clean way to schedule posts, upload media, list connected accounts, and fetch social analytics through {{PRODUCT_NAME}}. Install the skill once, add your access key to the agent environment, and ask for a concrete outcome. The agent finds the right connected accounts, prepares the payload, and creates the scheduled post for review in {{DASHBOARD_URL}}.

What are agent skills?

Agent skills are small instruction packs for coding agents. Instead of explaining an API, its authentication rules, payload shapes, and common workflows in every prompt, you install a skill once. The agent can then read the skill when the task calls for it and follow the intended workflow.

With {{PRODUCT_NAME}} agent skills, the agent learns how to authenticate with your access key through {{ACCESS_KEY_ENV_VAR_NAME}}, list connected social accounts from {{CONNECTED_ACCOUNTS}}, create and schedule posts for {{TARGET_PLATFORMS}}, upload or prepare media from {{MEDIA_ASSETS}} before publishing, inspect, reschedule, or cancel scheduled posts, and fetch analytics for published posts.

You stay in control. The skill gives the agent the workflow; your access key and prompt decide what it can actually do. The agent does not need to hand-roll separate logic for X, LinkedIn, Instagram, TikTok, YouTube, Pinterest, Facebook, Threads, and Bluesky every time it wants to post something.

How to use them

Install the skills source in your agent environment with the command shown in Install agent skills. Then add your access key wherever your agent reads environment variables, using {{ACCESS_KEY_ENV_VAR_NAME}}={{ACCESS_KEY_VALUE}}. Never commit {{ACCESS_KEY_VALUE}} to source control.

After setup, ask for a concrete outcome: Schedule a post for {{SCHEDULE_TIME}} on {{TARGET_PLATFORMS}} announcing {{POST_CONTENT}}. Prepare any required media from {{MEDIA_ASSETS}} and leave the scheduled post pending review. The agent uses the skill to find the right connected accounts, prepare the payload, and create the scheduled post through {{PRODUCT_NAME}}.

Scheduling with human review

The safest pattern is scheduling, not instant publishing. Ask the agent to turn {{SOURCE_MATERIAL}} into posts for {{TARGET_PLATFORMS}}, schedule them for {{SCHEDULE_TIME}}, and leave them in {{DASHBOARD_URL}} for review. You can edit, cancel, or let them go live.

That review step matters. Agents are useful, but they should not fire off production posts with no human in the loop unless {{REVIEW_POLICY}} documents a trusted workflow. Keep {{REVIEW_STATUS}} as pending review until you approve each post.

Skills or MCP?

Use agent skills when your coding agent works well with skill files and repeatable workflows. Use an MCP-compatible path when you want an MCP-compatible client, such as Claude Desktop, Cursor, or another Streamable HTTP client, to call {{PRODUCT_NAME}} tools directly. Both paths use the same {{PRODUCT_NAME}} API underneath. Skills fit coding-agent environments; MCP fits tool-calling clients. See Skills or MCP decision criteria for the side-by-side comparison.

Try it

Install the skills source, add your access key to your agent environment, and ask it to schedule a post: Schedule this for {{SCHEDULE_TIME}} on {{TARGET_PLATFORMS}}: {{POST_CONTENT}}. Use media from {{MEDIA_ASSETS}}. Leave the scheduled post in {{DASHBOARD_URL}} with review status {{REVIEW_STATUS}}. The agent discovers the connected accounts in {{CONNECTED_ACCOUNTS}}, creates the scheduled post, and leaves the queue where you can inspect it. Documentation is at {{DOCS_URL}} and MCP documentation is at {{MCP_DOCS_URL}}.

Failure checks

If authentication fails, check {{ACCESS_KEY_ENV_VAR_NAME}} in the agent environment. If no account is found for a platform, check that {{CONNECTED_ACCOUNTS}} includes an account for each platform in {{TARGET_PLATFORMS}} and that each platform is in {{SUPPORTED_PLATFORMS}}. If media is rejected, check {{MEDIA_ASSETS}} for the required file and format. If timing is wrong, check that {{SCHEDULE_TIME}} includes date, time, and timezone. If a post is ready to publish without review, check {{REVIEW_STATUS}} and hold it in {{DASHBOARD_URL}} until you approve it. See the failure matrix for the full checklist.

Agent skills field map
FieldValuePurpose
Skills source{{SKILLS_REPO}}Install source for agent skills
Access key variable{{ACCESS_KEY_ENV_VAR_NAME}}Environment variable the agent reads
API origin{{API_BASE_URL}}Origin for API calls; endpoint paths stay exact
Connected accounts{{CONNECTED_ACCOUNTS}}Accounts available for posting
Target platforms{{TARGET_PLATFORMS}}Platforms selected for this post
Post content{{POST_CONTENT}}Copy to schedule
Media assets{{MEDIA_ASSETS}}Files to upload before publishing
Schedule time{{SCHEDULE_TIME}}Date, time, and timezone
Review status{{REVIEW_STATUS}}Pending review until approved

Values are supplied inline as {{VARIABLE}} tokens.

Skills or MCP decision criteria
CriterionSkillsMCP
Best fitCoding agents that read skill filesTool-calling clients that support MCP
Setup input{{SKILLS_REPO}} and {{ACCESS_KEY_ENV_VAR_NAME}}{{MCP_DOCS_URL}} and MCP configuration
Workflow styleRepeatable, instruction-led workflowsDirect tool calls from the client
Example clients{{AGENT_ENVIRONMENT}}Claude Desktop, Cursor, Streamable HTTP clients

Decision criteria for choosing an integration path.

Failure matrix
Failure or riskLikely causeCheckAction
Agent cannot authenticateMissing or wrong access key{{ACCESS_KEY_ENV_VAR_NAME}} is setSet the key; never commit it
No account found for a platformAccount not connected{{CONNECTED_ACCOUNTS}} covers {{TARGET_PLATFORMS}}Connect the account or remove the platform
Platform not supportedPlatform outside {{SUPPORTED_PLATFORMS}}Compare target list against supported listRemove the unsupported platform
Media rejected or missingAsset not supplied or wrong format{{MEDIA_ASSETS}} is completeSupply or replace the asset
Wrong schedule timeTimezone or date ambiguity{{SCHEDULE_TIME}} includes timezoneCorrect the time and reschedule
Unreviewed instant publishingReview step skipped{{REVIEW_STATUS}} is not approvedCancel or hold the post in {{DASHBOARD_URL}}

Safeguard matrix for scheduling with review.

Frequently asked questions

What are agent skills?

Agent skills are small instruction packs for coding agents. Instead of repeating API, authentication, payload, and workflow instructions in every prompt, you install a skill once and the agent reads it when a relevant task arises. The skill does not replace your access key, connected accounts, or review decision.

What do I need before an agent can act?

You need {{SKILLS_REPO}}, {{ACCESS_KEY_ENV_VAR_NAME}}, {{API_BASE_URL}}, {{CONNECTED_ACCOUNTS}}, {{TARGET_PLATFORMS}}, {{POST_CONTENT}}, {{MEDIA_ASSETS}}, {{SCHEDULE_TIME}}, and {{REVIEW_STATUS}} before an agent can act. Each value is supplied inline where it is used.

Should I use skills or MCP?

Use skills when your coding agent works well with skill files and repeatable workflows. Use an MCP-compatible path when your client supports MCP tool calling. The choice depends on your client environment.

What is the safest publishing pattern?

Schedule first, then review. Leave scheduled posts in {{DASHBOARD_URL}} for you to edit, cancel, or approve before publishing. Instant publishing without review stays disabled unless {{REVIEW_POLICY}} documents a trusted workflow.

What should I check if scheduling fails?

Check {{ACCESS_KEY_ENV_VAR_NAME}}, confirm at least one account in {{CONNECTED_ACCOUNTS}} is connected for each platform in {{TARGET_PLATFORMS}}, supply any required media in {{MEDIA_ASSETS}}, and confirm {{REVIEW_STATUS}} before scheduling.

Install agent skills
npx skills add {{SKILLS_REPO}} --all