Class: RailsConsoleAi::Channel::Slack
- Defined in:
- lib/rails_console_ai/channel/slack.rb
Constant Summary collapse
- ANSI_REGEX =
/\e\[[0-9;]*m/
Instance Method Summary collapse
-
#add_guidance(text) ⇒ Object
Guidance is broadcast to both the main-engine queue and the sub-agent queue so a steering message arriving during a sub-agent run is seen by both layers (sub-agent reacts immediately; main engine reacts after delegate_task returns).
- #cancel! ⇒ Object
- #cancelled? ⇒ Boolean
- #confirm(_text) ⇒ Object
- #console_capture_string ⇒ Object
- #display(text) ⇒ Object
- #display_code(code) ⇒ Object
- #display_error(text) ⇒ Object
- #display_result(_result) ⇒ Object
- #display_result_output(output) ⇒ Object
- #display_status(text) ⇒ Object
- #display_thinking(text) ⇒ Object
- #display_tool_call(text) ⇒ Object
- #display_warning(text) ⇒ Object
- #drain_guidance(scope: :main) ⇒ Object
-
#initialize(slack_bot:, channel_id:, thread_ts:, user_name: nil) ⇒ Slack
constructor
A new instance of Slack.
- #log_input(text) ⇒ Object
- #mode ⇒ Object
- #pending_guidance?(scope: :main) ⇒ Boolean
- #prompt(text) ⇒ Object
-
#receive_reply(text) ⇒ Object
Called by SlackBot when a thread reply arrives.
- #supports_danger? ⇒ Boolean
- #supports_editing? ⇒ Boolean
- #system_instructions ⇒ Object
- #user_identity ⇒ Object
- #wrap_llm_call(&block) ⇒ Object
Methods inherited from Base
Constructor Details
#initialize(slack_bot:, channel_id:, thread_ts:, user_name: nil) ⇒ Slack
Returns a new instance of Slack.
8 9 10 11 12 13 14 15 16 17 18 19 20 |
# File 'lib/rails_console_ai/channel/slack.rb', line 8 def initialize(slack_bot:, channel_id:, thread_ts:, user_name: nil) @slack_bot = slack_bot @channel_id = channel_id @thread_ts = thread_ts @user_name = user_name @reply_queue = Queue.new @guidance_main = [] @guidance_sub = [] @guidance_mutex = Mutex.new @cancelled = false @log_prefix = "[#{@channel_id}/#{@thread_ts}] @#{@user_name}" @output_log = StringIO.new end |
Instance Method Details
#add_guidance(text) ⇒ Object
Guidance is broadcast to both the main-engine queue and the sub-agent queue so a steering message arriving during a sub-agent run is seen by both layers (sub-agent reacts immediately; main engine reacts after delegate_task returns).
33 34 35 36 37 38 |
# File 'lib/rails_console_ai/channel/slack.rb', line 33 def add_guidance(text) @guidance_mutex.synchronize do @guidance_main << text @guidance_sub << text end end |
#cancel! ⇒ Object
22 23 24 25 26 27 28 |
# File 'lib/rails_console_ai/channel/slack.rb', line 22 def cancel! @cancelled = true @guidance_mutex.synchronize do @guidance_main.clear @guidance_sub.clear end end |
#cancelled? ⇒ Boolean
56 57 58 |
# File 'lib/rails_console_ai/channel/slack.rb', line 56 def cancelled? @cancelled end |
#confirm(_text) ⇒ Object
120 121 122 |
# File 'lib/rails_console_ai/channel/slack.rb', line 120 def confirm(_text) 'y' end |
#console_capture_string ⇒ Object
188 189 190 |
# File 'lib/rails_console_ai/channel/slack.rb', line 188 def console_capture_string @output_log.string end |
#display(text) ⇒ Object
60 61 62 |
# File 'lib/rails_console_ai/channel/slack.rb', line 60 def display(text) post(strip_ansi(text)) end |
#display_code(code) ⇒ Object
95 96 97 98 99 100 101 |
# File 'lib/rails_console_ai/channel/slack.rb', line 95 def display_code(code) # Don't post raw code/plan steps to Slack — non-technical users don't need to see Ruby # But do log to STDOUT so server logs show what was generated/executed @output_log.write("# Generated code:\n#{code}\n") STDOUT.puts "#{@log_prefix} (code)" code.each_line { |line| STDOUT.puts "#{@log_prefix} (code) #{line.rstrip}" } end |
#display_error(text) ⇒ Object
86 87 88 |
# File 'lib/rails_console_ai/channel/slack.rb', line 86 def display_error(text) post(":x: #{strip_ansi(text)}") end |
#display_result(_result) ⇒ Object
110 111 112 113 |
# File 'lib/rails_console_ai/channel/slack.rb', line 110 def display_result(_result) # Don't post raw return values to Slack — the LLM formats output via puts nil end |
#display_result_output(output) ⇒ Object
103 104 105 106 107 108 |
# File 'lib/rails_console_ai/channel/slack.rb', line 103 def display_result_output(output) text = strip_ansi(output).strip return if text.empty? text = text[0, 3000] + "\n... (truncated)" if text.length > 3000 post("```#{text}```") end |
#display_status(text) ⇒ Object
70 71 72 73 74 75 76 77 78 79 80 |
# File 'lib/rails_console_ai/channel/slack.rb', line 70 def display_status(text) stripped = strip_ansi(text).strip return if stripped.empty? if stripped =~ /\AThinking\.\.\.|\AAttempting to fix|\ACancelled|\A_session:/ post(stripped) else @output_log.write("#{stripped}\n") log_prefixed("(status)", stripped) end end |
#display_thinking(text) ⇒ Object
64 65 66 67 68 |
# File 'lib/rails_console_ai/channel/slack.rb', line 64 def display_thinking(text) stripped = strip_ansi(text).strip return if stripped.empty? post(stripped) end |
#display_tool_call(text) ⇒ Object
90 91 92 93 |
# File 'lib/rails_console_ai/channel/slack.rb', line 90 def display_tool_call(text) @output_log.write("-> #{text}\n") log_prefixed("->", text) end |
#display_warning(text) ⇒ Object
82 83 84 |
# File 'lib/rails_console_ai/channel/slack.rb', line 82 def display_warning(text) post(":warning: #{strip_ansi(text)}") end |
#drain_guidance(scope: :main) ⇒ Object
40 41 42 43 44 45 46 47 |
# File 'lib/rails_console_ai/channel/slack.rb', line 40 def drain_guidance(scope: :main) @guidance_mutex.synchronize do arr = scope == :sub ? @guidance_sub : @guidance_main pending = arr.dup arr.clear pending end end |
#log_input(text) ⇒ Object
178 179 180 |
# File 'lib/rails_console_ai/channel/slack.rb', line 178 def log_input(text) @output_log.write("@#{@user_name}: #{text}\n") end |
#mode ⇒ Object
128 129 130 |
# File 'lib/rails_console_ai/channel/slack.rb', line 128 def mode 'slack' end |
#pending_guidance?(scope: :main) ⇒ Boolean
49 50 51 52 53 54 |
# File 'lib/rails_console_ai/channel/slack.rb', line 49 def pending_guidance?(scope: :main) @guidance_mutex.synchronize do arr = scope == :sub ? @guidance_sub : @guidance_main !arr.empty? end end |
#prompt(text) ⇒ Object
115 116 117 118 |
# File 'lib/rails_console_ai/channel/slack.rb', line 115 def prompt(text) post(strip_ansi(text)) @reply_queue.pop end |
#receive_reply(text) ⇒ Object
Called by SlackBot when a thread reply arrives
183 184 185 186 |
# File 'lib/rails_console_ai/channel/slack.rb', line 183 def receive_reply(text) @output_log.write("@#{@user_name}: #{text}\n") @reply_queue.push(text) end |
#supports_danger? ⇒ Boolean
132 133 134 |
# File 'lib/rails_console_ai/channel/slack.rb', line 132 def supports_danger? false end |
#supports_editing? ⇒ Boolean
136 137 138 |
# File 'lib/rails_console_ai/channel/slack.rb', line 136 def supports_editing? false end |
#system_instructions ⇒ Object
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 |
# File 'lib/rails_console_ai/channel/slack.rb', line 144 def system_instructions <<~INSTRUCTIONS.strip ## Slack Channel You are responding in a Slack thread. ## Code Execution - ALWAYS use the `execute_code` tool to run Ruby code. Do NOT put code in markdown code fences expecting it to be executed — code fences are display-only in Slack. - Use `execute_code` for simple queries, and `execute_plan` for multi-step operations. - If the user asks you to provide code they can run later, put it in a code fence in your text response (it will be displayed but not executed). ## Formatting - Slack does NOT support markdown tables. For tabular data, use `puts` to print a plain-text table inside a code block. Use fixed-width columns with padding so columns align. Example format: ``` ID Name Email 123 John Smith john@example.com 456 Jane Doe jane@example.com ``` - Use `puts` with formatted output instead of returning arrays or hashes. - Never return raw Ruby objects — always present data in a human-readable way. - The output of `puts` in your code is automatically shown to the user. Do NOT repeat or re-display data that your code already printed via `puts`. Just add a brief summary after (e.g. "10 events found" or "Let me know if you need more detail"). - Do not offer to make changes or take actions on behalf of the user. Only report findings. - This is a live production database — other processes, users, and background jobs are constantly changing data. Never assume results will be the same as a previous query. Always re-run queries when asked, even if you just ran the same one. INSTRUCTIONS end |
#user_identity ⇒ Object
124 125 126 |
# File 'lib/rails_console_ai/channel/slack.rb', line 124 def user_identity @user_name end |
#wrap_llm_call(&block) ⇒ Object
140 141 142 |
# File 'lib/rails_console_ai/channel/slack.rb', line 140 def wrap_llm_call(&block) yield end |