slack-max-api-mcp 1.0.9 → 1.0.10

package/.env.example

@@ -41,9 +41,19 @@ SLACK_CLIENT_SECRET=
 # SLACK_API_BASE_URL=https://slack.com/api
 
 # Tool exposure settings
-# SLACK_TOOL_EXPOSURE_MODE=smart
-# SLACK_SMART_COMPAT_CORE_TOOLS=true
+# SLACK_TOOL_EXPOSURE_MODE=operations
+# SLACK_SMART_COMPAT_CORE_TOOLS=true   # developer mode only; keep raw core tools visible
 # SLACK_ENABLE_METHOD_TOOLS=false
 # SLACK_METHOD_TOOL_PREFIX=slack_method
 # SLACK_MAX_METHOD_TOOLS=50
 # SLACK_CATALOG_PATH=C:\code_test\test22\data\slack-catalog.json
+
+# Governance / operations settings
+# SLACK_OPERATIONS_CONFIG_PATH=C:\code_test\test22\config\operations.json
+# SLACK_OPERATIONS_STATE_PATH=C:\Users\<you>\.slack-max-api-mcp\operations-state.json
+# SLACK_ENABLE_AUDIT_LOG=true
+# SLACK_AUDIT_LOG_PATH=C:\Users\<you>\.slack-max-api-mcp\audit.log
+# SLACK_METHOD_ALLOWLIST=chat.postMessage,conversations.history
+# SLACK_METHOD_DENYLIST=users.deletePhoto
+# SLACK_METHOD_ALLOW_PREFIXES=chat.,conversations.
+# SLACK_METHOD_DENY_PREFIXES=admin.

package/README.md

@@ -1,6 +1,6 @@
 # Slack Max API MCP
 
-Slack Web API MCP server for Codex/Claude Code over `stdio`.
+Operations-first Slack MCP server for Codex/Claude Code over `stdio`.
 
 - Package: `slack-max-api-mcp`
 - Runtime: pure CLI MCP over stdio
@@ -9,16 +9,86 @@ Slack Web API MCP server for Codex/Claude Code over `stdio`.
 
 ## Tool exposure
 
-- `smart` (default): 16 tools (`gateway_*` router names + core tools)
+- `operations` (default): 14 operations-first tools only
+- `developer`: operations tools + `gateway_*` + core Slack API tools
 - `legacy`: fixed core tools + optional catalog method tools
 
 Environment:
 
-- `SLACK_TOOL_EXPOSURE_MODE=smart|legacy`
+- `SLACK_TOOL_EXPOSURE_MODE=operations|developer|legacy`
 - `SLACK_SMART_COMPAT_CORE_TOOLS=true|false`
 - `SLACK_ENABLE_METHOD_TOOLS=true|false`
 - `SLACK_MAX_METHOD_TOOLS=<number>`
 
+`smart` is still accepted as an alias for `developer`.
+
+## Operations-first tools
+
+The default surface is now operations-first and keeps raw API wrappers out of the main tool list:
+
+- `ops_policy_info`: runtime policy/audit guardrails
+- `ops_playbook_list`: built-in operations playbooks
+- `ops_state_overview`: inspect local operations state (`incidents`, `digests`, `broadcasts`, `followups`)
+- `ops_incident_create`: create + optionally announce a tracked incident
+- `ops_incident_update`: persist a status change and optionally post a thread update
+- `ops_incident_close`: close a tracked incident with stored resolution
+- `ops_broadcast_prepare`: prepare and store a broadcast draft before sending
+- `ops_playbook_run`: run standardized workflows (`incident_open`, `support_digest`, `release_broadcast`)
+- `ops_channel_snapshot`: activity/participant/thread snapshot for a channel
+- `ops_unanswered_threads`: find stale or unanswered question-like threads
+- `ops_sla_breach_scan`: detect SLA breach threads across multiple channels
+- `ops_sla_followup`: auto follow-up replies for SLA breaches with duplicate-suppression state
+- `ops_broadcast_message`: send a prepared draft or direct operational announcement
+- `ops_audit_log_read`: inspect local JSONL audit logs
+
+These tools let teams run repeatable Slack operations without rebuilding multi-step API call chains, and they persist local operational state to make incidents/broadcasts/followups first-class records.
+
+## Config and state
+
+- Operations config file: `config/operations.json`
+- Override path: `SLACK_OPERATIONS_CONFIG_PATH=<path>`
+- Local state path: `~/.slack-max-api-mcp/operations-state.json`
+- Override path: `SLACK_OPERATIONS_STATE_PATH=<path>`
+
+The operations config controls incident templates, digest defaults, broadcast templates, and follow-up suppression windows.
+
+### Playbook examples
+
+```text
+Codex, run ops_playbook_run with playbook=incident_open on #incident-war-room.
+title is "DB Latency Spike", severity is "sev2", owner is "@oncall-db", dry_run=true.
+```
+
+```text
+Codex, run ops_playbook_run with playbook=support_digest for channels #support-kor and #support-global.
+lookback_hours=24, sla_minutes=60, report_channel=#support-ops, dry_run=true.
+```
+
+```text
+Codex, run ops_sla_followup for channels #support-kor and #support-global.
+sla_minutes=90, lookback_hours=24, max_messages=20, dry_run=true.
+```
+
+```text
+Codex, run ops_incident_create on #incident-war-room.
+title is "API Error Spike", summary is "5xx rate above threshold", owner is "@oncall-api", dry_run=true.
+```
+
+```text
+Codex, run ops_state_overview with collection=incidents and limit=10.
+```
+
+## Governance settings
+
+- `SLACK_ENABLE_AUDIT_LOG=true|false` (default: `true`)
+- `SLACK_AUDIT_LOG_PATH=<path>` (default: `~/.slack-max-api-mcp/audit.log`)
+- `SLACK_METHOD_ALLOWLIST=chat.postMessage,conversations.history`
+- `SLACK_METHOD_DENYLIST=users.deletePhoto`
+- `SLACK_METHOD_ALLOW_PREFIXES=chat.,conversations.`
+- `SLACK_METHOD_DENY_PREFIXES=admin.`
+
+If allowlist/allow-prefix is set, methods outside that policy are blocked.
+
 ## Install
 
 ```powershell

package/config/operations.json

@@ -0,0 +1,75 @@
+{
+  "playbooks": {
+    "incident_open": {
+      "summary": "Post an incident opening message with a standard operations template.",
+      "required_inputs": ["channel", "title"],
+      "supports_dry_run": true
+    },
+    "support_digest": {
+      "summary": "Build and optionally publish a support digest from channel activity.",
+      "required_inputs": ["channels"],
+      "supports_dry_run": true
+    },
+    "release_broadcast": {
+      "summary": "Broadcast a release announcement to multiple channels.",
+      "required_inputs": ["channels", "title", "summary"],
+      "supports_dry_run": true
+    }
+  },
+  "incidents": {
+    "default_channel": "",
+    "default_owner": "TBD",
+    "default_severity": "sev2",
+    "update_interval_minutes": 15,
+    "open_template": [
+      ":rotating_light: Incident Open - {{title}}",
+      "Severity: {{severity_upper}}",
+      "Owner: {{owner}}",
+      "Summary: {{summary}}",
+      "{{details_line}}",
+      "Next update: {{next_update_text}}"
+    ],
+    "update_template": [
+      ":information_source: Incident Update - {{title}}",
+      "Status: {{status_upper}}",
+      "Severity: {{severity_upper}}",
+      "Owner: {{owner}}",
+      "Summary: {{summary}}",
+      "{{details_line}}",
+      "Next update: {{next_update_text}}"
+    ],
+    "close_template": [
+      ":white_check_mark: Incident Resolved - {{title}}",
+      "Severity: {{severity_upper}}",
+      "Owner: {{owner}}",
+      "Resolution: {{resolution}}",
+      "{{details_line}}"
+    ]
+  },
+  "support_digest": {
+    "default_lookback_hours": 24,
+    "default_sla_minutes": 60,
+    "default_max_threads": 10,
+    "header_template": "Support Digest ({{lookback_hours}}h)",
+    "sla_template": "SLA threshold: {{sla_minutes}}m"
+  },
+  "broadcasts": {
+    "default_template": "release_default",
+    "default_mrkdwn": true,
+    "templates": {
+      "release_default": [
+        ":rocket: Release - {{title}}",
+        "{{summary}}",
+        "{{details}}"
+      ]
+    }
+  },
+  "followups": {
+    "default_sla_minutes": 60,
+    "default_lookback_hours": 24,
+    "default_max_threads_per_channel": 20,
+    "default_max_messages": 30,
+    "suppress_hours": 6,
+    "reminder_template": "Friendly reminder: this thread appears to be pending for more than {{sla_minutes}} minutes. Please provide an update."
+  }
+}

package/package.json

@@ -1,13 +1,14 @@
 {
   "name": "slack-max-api-mcp",
-  "version": "1.0.9",
-  "description": "Slack MCP server (stdio) for Codex and Claude Code",
+  "version": "1.0.10",
+  "description": "Operations-first Slack MCP server (stdio) for Codex and Claude Code",
   "main": "src/slack-mcp-server.js",
   "bin": {
     "slack-max-api-mcp": "src/slack-mcp-server.js"
   },
   "files": [
     "src",
+    "config",
     "data/slack-catalog.json",
     "README.md",
     ".env.example"