AGENT SKILL
pm-freshchat
Freshchat connector knowledge and safe action guide.
- License 许可
- AGPL-3.0
- 最近更新
- 2026-09-21
- Stars
- ★ 7
- Forks
- 0
01 前言 Front Matter
- name
- pm-freshchat
- description
- Freshchat connector knowledge and safe action guide.
03 SKILL.md 正文
pm-freshchat
Purpose
Reads Freshchat account, user, conversation, agent, group, channel, role, outbound, report, metrics, and business-hours data through the Freshchat v2 REST API; writes Freshchat users, conversations, agents, outbound messages, reports, and CSAT ratings.
Icon
- id: pm-sample
- asset: icons/pm-sample.svg
- source: polymetrics
- review_status: polymetrics
- review_url: https://github.com/polymetrics-ai/cli
Capabilities
- check=true catalog=true read=true write=true query=false
- Integration type: api
Authentication
- Use pm credentials add with --from-env or --value-stdin for secret fields.
Configuration
- agent_id
- agents_availability_status
- agents_groups
- agents_is_deactivated
- agents_sort_by
- agents_sort_order
- base_url (required)
- business_hours_group_id
- channels_locale
- conversation_id
- conversation_messages_from_time
- groups_sort_by
- groups_sort_order
- metrics_aggregator
- metrics_end
- metrics_filter_by
- metrics_group_by
- metrics_interval
- metrics_metric
- metrics_start
- metrics_summary
- mode
- outbound_request_id
- report_id
- report_status
- user_id
- users_created_from
- users_created_to
- users_email
- users_first_name
- users_last_name
- users_phone_no
- users_reference_id
- users_updated_from
- users_updated_to
- api_key (secret) (required)
ETL Streams
- account_configuration:
- primary key: account_id
- fields: account_domain(string), account_id(integer), app_id(string), bundle_id(integer), bundle_type(string), datacenter(string), organisation_domain(string), organisation_id(integer), plan_type(string)
- agents:
- primary key: id
- cursor: updated_time
- fields: avatar(object), biography(string), created_time(string), email(string), first_name(string), groups(array), id(string), is_deactivated(boolean), is_deleted(boolean), last_name(string), role_id(string), social_profiles(array), updated_time(string)
- agent_details:
- primary key: id
- cursor: updated_time
- fields: avatar(object), biography(string), created_time(string), email(string), first_name(string), groups(array), id(string), is_deactivated(boolean), is_deleted(boolean), last_name(string), role_id(string), social_profiles(array), updated_time(string)
- agent_statuses:
- primary key: id
- fields: enabled(boolean), id(string), name(string), type(string)
- users:
- primary key: id
- cursor: updated_time
- fields: avatar(object), created_time(string), email(string), first_name(string), id(string), last_name(string), phone(string), properties(array), reference_id(string), restore_id(string), updated_time(string)
- user_details:
- primary key: id
- cursor: updated_time
- fields: avatar(object), created_time(string), email(string), first_name(string), id(string), last_name(string), phone(string), properties(array), reference_id(string), restore_id(string), updated_time(string)
- user_conversations:
- primary key: id
- fields: app_id(string), assigned_agent_id(string), assigned_group_id(string), channel_id(string), created_time(string), id(string), messages(array), priority(string), properties(object), status(string), updated_time(string), user_id(string)
- conversation_detail:
- primary key: id
- fields: app_id(string), assigned_agent_id(string), assigned_group_id(string), channel_id(string), created_time(string), id(string), messages(array), priority(string), properties(object), status(string), updated_time(string), user_id(string)
- conversation_messages:
- primary key: id
- fields: actor_id(string), actor_type(string), app_id(string), conversation_id(string), created_time(string), id(string), message_parts(array), message_type(string), updated_time(string)
- conversation_fields:
- primary key: name
- fields: choices(array), label(string), name(string), required(boolean), type(string)
- groups:
- primary key: id
- fields: created_time(string), description(string), id(string), name(string), routing_type(string), updated_time(string)
- channels:
- primary key: id
- cursor: updated_time
- fields: created_time(string), enabled(boolean), icon(object), id(string), locale(string), name(string), public(boolean), tags(array), updated_time(string), welcome_message(object)
- roles:
- primary key: id
- fields: description(string), id(string), name(string), role(string)
- outbound_messages:
- primary key: id
- fields: created_time(string), from(object), id(string), provider(string), request_id(string), status(string), template(object), to(object), updated_time(string)
- report_status:
- primary key: id
- fields: id(string), interval(string), link(object), links(array), status(string)
- historical_metrics:
- primary key: metric_type
- fields: aggregator(string), data(array), end(string), filters(object), interval(string), metric_type(string), metrics(array), start(string)
- instant_metrics:
- primary key: metric_type
- fields: aggregator(string), data(array), end(string), filters(object), interval(string), metric_type(string), metrics(array), start(string)
- business_hours_status:
- primary key: group_id
- fields: business_hours_id(string), group_id(string), timezone(string), within_business_hours(boolean)
Sync Modes
- ETL sync modes: full_refresh_append, full_refresh_overwrite, full_refresh_overwrite_deduped, incremental_append, incremental_append_deduped
Reverse ETL Actions
- create_user:
- endpoint: POST /users
- risk: creates a Freshchat user/contact visible to agents
- update_user:
- endpoint: PUT /users/{{ record.user_id }}
- required fields: user_id
- risk: updates an existing Freshchat user/contact
- delete_user:
- endpoint: DELETE /users/{{ record.user_id }}
- required fields: user_id
- risk: deletes a Freshchat user/contact; destructive and idempotent for configured missing statuses
- create_conversation:
- endpoint: POST /conversations
- risk: creates a Freshchat conversation
- update_conversation:
- endpoint: PUT /conversations/{{ record.conversation_id }}
- required fields: conversation_id
- risk: updates routing, status, or properties on an existing Freshchat conversation
- send_conversation_message:
- endpoint: POST /conversations/{{ record.conversation_id }}/messages
- required fields: conversation_id
- risk: sends a message into an existing Freshchat conversation
- create_agent:
- endpoint: POST /agents
- required fields: email
- risk: creates a Freshchat admin/agent account; requires typed destructive confirmation because this is an administrative user-management action
- update_agent:
- endpoint: PUT /agents/{{ record.agent_id }}
- required fields: agent_id
- risk: updates a Freshchat admin/agent account; requires typed destructive confirmation because this mutates administrative access metadata
- update_agent_status:
- endpoint: PATCH /agents/{{ record.agent_id }}
- required fields: agent_id, status
- risk: updates a Freshchat agent availability status; requires typed destructive confirmation for administrative agent-state mutation
- delete_agent:
- endpoint: DELETE /agents/{{ record.agent_id }}
- required fields: agent_id
- risk: deletes a Freshchat admin/agent account; destructive and idempotent for configured missing statuses
- send_outbound_whatsapp_message:
- endpoint: POST /outbound-messages/whatsapp
- risk: sends an outbound WhatsApp message through Freshchat
- extract_report:
- endpoint: POST /reports/raw
- required fields: start, end, event, format
- risk: requests generation of a Freshchat raw report extract
- create_csat_rating:
- endpoint: POST /csat/{{ record.conversation_id }}
- required fields: conversation_id
- risk: creates a CSAT rating for a Freshchat conversation
Security
- read risk: external Freshchat API reads of account metadata, users, conversations, messages, agents, groups, channels, roles, outbound messages, reports, metrics, and business-hours status
- write risk: creates, updates, or deletes Freshchat users, conversations, agents, outbound messages, reports, and CSAT ratings
- approval: reverse ETL writes require plan preview and approval token
- Never pass secret values in chat, shell arguments, logs, docs, or JSON output.
Command Surface
- Read Freshchat data and safely plan Freshchat write actions without raw provider escape hatches.
- Usage: pm freshchat <account|agents|users|conversations|groups|channels|roles|outbound|reports|metrics|business-hours|csat|direct|binary> <command> [flags]
- Source CLI: Freshchat API (Freshchat API docs ETag W/"26e4fd8b1fe01578eae1dbaff6b69224")
- Global flags:
- --credential (non-empty) (string): Credential profile name; never pass secret values as flags.: maps_to=config.credential
- --config (string_array): Connector config override as key=value; never pass secret values here.
- --json (boolean): Render machine-readable JSON output.
- --limit (integer): Maximum records to emit from stream-backed commands.
- --preview (boolean): Preview a reverse-ETL write without making a network mutation.
- --plan (string): Execute an approved reverse-ETL plan by id.
- --approval-token-stdin (boolean): Read the approval token as one bounded line from standard input.
- --confirm (string): Typed confirmation challenge for destructive/admin reverse-ETL writes.
- Freshchat account read commands
- account configuration - Read Freshchat account configuration. [intent=etl availability=implemented stream=account_configuration]
- Freshchat agent read/write/admin commands
- agents list - List Freshchat agents with optional documented filters. [intent=etl availability=implemented stream=agents]; flags: --is-deactivated, --groups (non-empty), --availability-status, --sort-order, --sort-by (non-empty)
- agents view - Read one Freshchat agent by configured agent id. [intent=etl availability=implemented stream=agent_details]; flags: --agent-id (non-empty)
- agents statuses - List Freshchat agent statuses. [intent=etl availability=implemented stream=agent_statuses]
- agents create - Plan creation of a Freshchat admin/agent account. [intent=reverse_etl availability=implemented write=create_agent]; approval: Destructive/admin reverse ETL writes require typed destructive confirmation plus plan -> preview -> explicit approval -> execute.; risk: Creates a Freshchat admin/agent account.; flags: --email (non-empty), --first-name (non-empty), --last-name (non-empty), --role-id (non-empty), --groups
- agents update - Plan update of a Freshchat admin/agent account. [intent=reverse_etl availability=implemented write=update_agent]; approval: Destructive/admin reverse ETL writes require typed destructive confirmation plus plan -> preview -> explicit approval -> execute.; risk: Updates a Freshchat admin/agent account.; flags: --agent-id (non-empty), --first-name (non-empty), --last-name (non-empty), --role-id (non-empty), --groups, --is-deactivated
- agents status update - Plan update of a Freshchat agent availability status. [intent=reverse_etl availability=implemented write=update_agent_status]; approval: Destructive/admin reverse ETL writes require typed destructive confirmation plus plan -> preview -> explicit approval -> execute.; risk: Updates a Freshchat agent availability status.; flags: --agent-id (non-empty), --status (non-empty)
- agents delete - Plan deletion of a Freshchat admin/agent account. [intent=reverse_etl availability=implemented write=delete_agent]; approval: Destructive/admin reverse ETL writes require typed destructive confirmation plus plan -> preview -> explicit approval -> execute.; risk: Deletes a Freshchat admin/agent account.; flags: --agent-id (non-empty)
- Freshchat user read/write commands
- users list - List Freshchat users with documented search filters. [intent=etl availability=implemented stream=users]; notes: Freshchat requires at least one user search filter for successful live /users responses; connector config and command metadata expose the documented filter set without a raw query escape hatch.; flags: --first-name (non-empty), --last-name (non-empty), --email (non-empty), --reference-id (non-empty), --phone-no (non-empty), --created-from (non-empty, format=date-time), --created-to (non-empty, format=date-time), --updated-from (non-empty, format=date-time), --updated-to (non-empty, format=date-time)
- users view - Read one Freshchat user by configured user id. [intent=etl availability=implemented stream=user_details]; flags: --user-id (non-empty)
- users conversations - List conversations for a Freshchat user. [intent=etl availability=implemented stream=user_conversations]; flags: --user-id (non-empty)
- users create - Plan creation of a Freshchat user/contact. [intent=reverse_etl availability=implemented write=create_user]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Creates a Freshchat user/contact visible to agents.; flags: --email (non-empty), --first-name (non-empty), --last-name (non-empty), --phone (non-empty), --reference-id (non-empty)
- users update - Plan update of a Freshchat user/contact. [intent=reverse_etl availability=implemented write=update_user]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Updates an existing Freshchat user/contact.; flags: --user-id (non-empty), --email (non-empty), --first-name (non-empty), --last-name (non-empty), --phone (non-empty), --reference-id (non-empty)
- users delete - Plan deletion of a Freshchat user/contact. [intent=reverse_etl availability=implemented write=delete_user]; approval: Destructive/admin reverse ETL writes require typed destructive confirmation plus plan -> preview -> explicit approval -> execute.; risk: Deletes a Freshchat user/contact.; flags: --user-id (non-empty)
- Freshchat conversation read/write commands
- conversations view - Read one Freshchat conversation. [intent=etl availability=implemented stream=conversation_detail]; flags: --conversation-id (non-empty)
- conversations messages - List messages in a Freshchat conversation. [intent=etl availability=implemented stream=conversation_messages]; flags: --conversation-id (non-empty), --from-time (non-empty, format=date-time)
- conversations fields - List Freshchat conversation property fields. [intent=etl availability=implemented stream=conversation_fields]
- conversations create - Plan creation of a Freshchat conversation. [intent=reverse_etl availability=implemented write=create_conversation]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Creates a Freshchat conversation.; flags: --user-id (non-empty), --channel-id (non-empty)
- conversations update - Plan update of a Freshchat conversation. [intent=reverse_etl availability=implemented write=update_conversation]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Updates status, routing, priority, or properties on a Freshchat conversation.; flags: --conversation-id (non-empty), --status (non-empty), --assigned-agent-id (non-empty), --assigned-group-id (non-empty), --priority (non-empty)
- conversations message send - Plan sending a message into a Freshchat conversation. [intent=reverse_etl availability=implemented write=send_conversation_message]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Sends a visible message into an existing Freshchat conversation.; flags: --conversation-id (non-empty), --actor-type (non-empty), --actor-id (non-empty)
- Freshchat group read commands
- groups list - List Freshchat groups. [intent=etl availability=implemented stream=groups]; flags: --sort-order, --sort-by (non-empty)
- Freshchat channel read commands
- channels list - List Freshchat channels/topics. [intent=etl availability=implemented stream=channels]; flags: --locale (non-empty)
- Freshchat role read commands
- roles list - List Freshchat roles. [intent=etl availability=implemented stream=roles]
- Freshchat outbound read/write commands
- outbound messages list - List Freshchat outbound messages. [intent=etl availability=implemented stream=outbound_messages]; flags: --request-id (non-empty)
- outbound whatsapp send - Plan sending an outbound WhatsApp message through Freshchat. [intent=reverse_etl availability=implemented write=send_outbound_whatsapp_message]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Sends an outbound WhatsApp message through Freshchat.; flags: --provider (non-empty)
- Freshchat report read/write commands
- reports status - Read Freshchat raw report status. [intent=etl availability=implemented stream=report_status]; flags: --report-id (non-empty), --status (non-empty)
- reports extract - Plan generation of a Freshchat raw report extract. [intent=reverse_etl availability=implemented write=extract_report]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Requests generation of a Freshchat raw report extract.; flags: --start (non-empty, format=date-time), --end (non-empty, format=date-time), --event (non-empty), --format
- Freshchat metrics read commands
- metrics historical - Read Freshchat historical metrics. [intent=etl availability=implemented stream=historical_metrics]; flags: --metric (non-empty), --start (non-empty, format=date-time), --end (non-empty, format=date-time), --group-by (non-empty), --filter-by (non-empty), --aggregator (non-empty), --interval (non-empty)
- metrics instant - Read Freshchat instant metrics. [intent=etl availability=implemented stream=instant_metrics]; flags: --metric (non-empty), --group-by (non-empty), --filter-by (non-empty), --summary (non-empty)
- Freshchat business-hours read commands
- business-hours status - Check whether a Freshchat group is within business hours. [intent=etl availability=implemented stream=business_hours_status]; flags: --group-id (non-empty)
- Freshchat CSAT write commands
- csat create - Plan creation of a Freshchat CSAT rating. [intent=reverse_etl availability=implemented write=create_csat_rating]; approval: reverse ETL writes require plan -> preview -> explicit approval -> execute.; risk: Creates a CSAT rating for a Freshchat conversation.; flags: --conversation-id (non-empty), --rating, --comment (non-empty)
- Planned bounded provider-search commands
- direct users fetch - Planned bounded subset fetch for up to 100 Freshchat user ids. [intent=direct_read availability=planned]; approval: none until #2985 provides executable provider_search/provider_query safety; no live call is available today.; risk: medium; notes: Blocked on #2985 typed provider-search/query foundation. The official request body is ids[] with a provider-documented maximum of 100 users; no raw body escape hatch is exposed.; flags: --ids, --page, --page-cursor
- Planned binary/multipart commands
- binary files upload - Planned Freshchat file upload. [intent=direct_write availability=planned]; approval: blocked until a typed binary/multipart plan -> preview -> approval -> execute contract exists.; risk: high; notes: Blocked binary/file operation. Official API accepts one multipart file with a documented 25 MB cap; this connector currently accepts no filesystem path or binary payload.; flags: --file (non-empty)
- binary images upload - Planned Freshchat image upload. [intent=direct_write availability=planned]; approval: blocked until a typed binary/multipart plan -> preview -> approval -> execute contract exists.; risk: high; notes: Blocked binary/file operation. Official API accepts multipart image input; this connector currently accepts no filesystem path or binary payload.; flags: --image (non-empty)
- Help topics:
- destructive-confirmation - Freshchat DELETE and admin/agent-management operations require typed destructive confirmation plus reverse ETL plan -> preview -> approval -> execute.
- provider-search - Freshchat POST /users/fetch remains planned/blocked on #2985 and is not exposed as a raw query/body command.
- binary-uploads - Freshchat file/image uploads remain planned/blocked until a typed binary/multipart safety contract exists.
Commands
Inspect as a manual
pm connectors inspect freshchat
Inspect as structured JSON
pm connectors inspect freshchat --json
Agent Rules
- Run pm connectors inspect freshchat before creating credentials or plans.
- Use --json only when the caller needs structured output; use the manual for human-readable guidance.
- Never ask the user to paste secret values into chat.
- For reverse ETL writes, create a plan, show the preview, wait for explicit approval, then run with the approval token.
04 同仓库更多 Skills
recipe-preview-approve-reverse-etl
Preview and approve a reverse ETL plan safely.
★ 7
recipe-github-prs-to-warehouse
Sync GitHub pull requests into the local warehouse.
★ 7
pm-zoom
Zoom connector knowledge and safe action guide.
★ 7
pm-zonka-feedback
Zonka Feedback connector knowledge and safe action guide.
★ 7
pm-zoho-invoice
Zoho Invoice connector knowledge and safe action guide.
★ 7
pm-zoho-inventory
Zoho Inventory connector knowledge and safe action guide.
★ 7