Planned
Title: Intermittent Threading Bug — Agent Responses Posted Outside Originating Thread
Severity: Medium-High | Intermittent / Recurring Agent ID: ff1379fe-51d1-4bd4-a870-e5ef6dc11d88 (Ask_Sam) Internal Ref: ANT-3200 Bug Description When Ask_Sam auto-triggers on a new message in #analytics (CCWKC4Z41), responses intermittently post as top-level channel messages instead of thread replies. A secondary related behavior: occasional duplicate responses — two messages fire on the same trigger within seconds. Expected: All auto-trigger responses post inside the originating thread. Actual: Response posts at channel level — unanchored, invisible to requestor, clutters channel. Frequency: Intermittent (~5–10% of triggers). Ongoing for several months. Previously reported; partially resolved; still recurring as of July 6, 2026 (screenshot attached showing ANT-3388 response orphaned at channel level). Business Impact Requestors miss their answers Channel noise accumulates with context-free orphaned messages Erodes stakeholder trust in agent reliability Directly impacts our internal N6 Operational Reliability KPI (95% target) Attachment: Screenshot provided showing orphaned response referencing ANT-3388 posted at top-level in #analytics.
Patrick Codrington 12 days ago
Planned
Title: Intermittent Threading Bug — Agent Responses Posted Outside Originating Thread
Severity: Medium-High | Intermittent / Recurring Agent ID: ff1379fe-51d1-4bd4-a870-e5ef6dc11d88 (Ask_Sam) Internal Ref: ANT-3200 Bug Description When Ask_Sam auto-triggers on a new message in #analytics (CCWKC4Z41), responses intermittently post as top-level channel messages instead of thread replies. A secondary related behavior: occasional duplicate responses — two messages fire on the same trigger within seconds. Expected: All auto-trigger responses post inside the originating thread. Actual: Response posts at channel level — unanchored, invisible to requestor, clutters channel. Frequency: Intermittent (~5–10% of triggers). Ongoing for several months. Previously reported; partially resolved; still recurring as of July 6, 2026 (screenshot attached showing ANT-3388 response orphaned at channel level). Business Impact Requestors miss their answers Channel noise accumulates with context-free orphaned messages Erodes stakeholder trust in agent reliability Directly impacts our internal N6 Operational Reliability KPI (95% target) Attachment: Screenshot provided showing orphaned response referencing ANT-3388 posted at top-level in #analytics.
Patrick Codrington 12 days ago
Multi-Agent Shared Protocol Blocks with API-Based Deployment
Summary: Organizations running multiple Runbear agents (e.g., Ask_Sam, Ask_HR, Ask_Eng) need a way to maintain shared behavioral rules — identity resolution, circuit breakers, access control, response protocols — across all agents without manually updating each one individually. Problem: Today, every system prompt update must be applied manually per agent via the Runbear UI. When a fix is validated on one agent (e.g., a user name resolution bug), replicating it to other agents requires: • Copy/pasting the updated block into each agent manually • No version control or audit trail at the deployment layer • High risk of drift between agents over time Requested Features: 1. Shared Protocol Library — A workspace-level repository of reusable prompt blocks (e.g., identity-resolution.md, circuit-breaker.md) that can be referenced by multiple agents 2. Agent Configuration API — A REST endpoint to programmatically read/write agent system prompts, enabling CI/CD integration (e.g., on GitLab MR merge → push updated shared blocks to all registered agents) 3. Inheritance Model — Allow agents to inherit base protocols + override with agent-specific config, similar to how.cursorrules works in Cursor Business Impact: Reduces admin overhead, eliminates agent drift, and enables enterprise-scale governance of AI agent behavior.
Patrick Codrington 2 months ago
Multi-Agent Shared Protocol Blocks with API-Based Deployment
Summary: Organizations running multiple Runbear agents (e.g., Ask_Sam, Ask_HR, Ask_Eng) need a way to maintain shared behavioral rules — identity resolution, circuit breakers, access control, response protocols — across all agents without manually updating each one individually. Problem: Today, every system prompt update must be applied manually per agent via the Runbear UI. When a fix is validated on one agent (e.g., a user name resolution bug), replicating it to other agents requires: • Copy/pasting the updated block into each agent manually • No version control or audit trail at the deployment layer • High risk of drift between agents over time Requested Features: 1. Shared Protocol Library — A workspace-level repository of reusable prompt blocks (e.g., identity-resolution.md, circuit-breaker.md) that can be referenced by multiple agents 2. Agent Configuration API — A REST endpoint to programmatically read/write agent system prompts, enabling CI/CD integration (e.g., on GitLab MR merge → push updated shared blocks to all registered agents) 3. Inheritance Model — Allow agents to inherit base protocols + override with agent-specific config, similar to how.cursorrules works in Cursor Business Impact: Reduces admin overhead, eliminates agent drift, and enables enterprise-scale governance of AI agent behavior.
Patrick Codrington 2 months ago
Completed
Allow cross-channel Slcak posting when channel is explicitly identified in prompt
Description: When a user references a specific Slack channel in their prompt (e.g., "post this update to #CCY8M69PZ"), Sam should be able to post to that channel even when the originating prompt comes from a different channel. Current behavior: Sam can only post to the channel where the session originated. If a user asks Sam to post to another channel, Sam acknowledges the request but fails silently or acknowledges it can't follow through — creating a trust gap. Expected behavior: If a target channel is explicitly named in the prompt, Sam should be able to post there (subject to appropriate permissions). Use case: Analytics team uses Sam across multiple channels. Users frequently ask Sam to post updates to originating Jira/project threads, pipeline channels, or stakeholder channels based on context — not just the channel where they're prompting.
Patrick Codrington 2 months ago
Completed
Allow cross-channel Slcak posting when channel is explicitly identified in prompt
Description: When a user references a specific Slack channel in their prompt (e.g., "post this update to #CCY8M69PZ"), Sam should be able to post to that channel even when the originating prompt comes from a different channel. Current behavior: Sam can only post to the channel where the session originated. If a user asks Sam to post to another channel, Sam acknowledges the request but fails silently or acknowledges it can't follow through — creating a trust gap. Expected behavior: If a target channel is explicitly named in the prompt, Sam should be able to post there (subject to appropriate permissions). Use case: Analytics team uses Sam across multiple channels. Users frequently ask Sam to post updates to originating Jira/project threads, pipeline channels, or stakeholder channels based on context — not just the channel where they're prompting.
Patrick Codrington 2 months ago
Capability for Subagent Creation for runbear_file_search (and other tasks)
Runbear Feature Request: Agent-to-Agent Invocation & KB Search Subagents Title: Enable Subagent Creation for Knowledge Base Search Delegation Priority: High Use Case: Token Budget Management & Scalability Submitted by: Minted Analytics Team (Ask_Sam Agent) *** Problem Statement Our Ask_Sam agent (ff1379fe-51d1-4bd4-a870-e5ef6dc11d88) frequently hits the 1M token context limit due to: 1. 144 active tools consuming ~2,400 tokens per turn just for tool definitions 2. runbear_file_search results returning 5,000-15,000 tokens per query 3. Long-term memory (LTM) accumulating ~30,000 tokens of historical learnings 4. Multi-turn conversations requiring full context retention Current Impact: • Conversations terminated prematurely with "token limit exceeded" errors • User experience disrupted mid-analysis • Complex requests cannot be completed *** Requested Feature: Agent-to-Agent Invocation Core Capability: runbear_invoke_agent( agent_id: str,# Specialized subagent ID task: str,# Scoped instruction max_context_return: int,# Token limit for summary response pass_context: bool = False# Whether to share parent context ) Returns: • Compressed summary (user-defined token limit) • Full results remain in subagent's context • Parent agent receives only the synthesized response *** Proposed Architecture Pattern 1: Knowledge Base Search Specialist Ask_Sam (Parent Agent) ↓ Invokes KB_Search_Specialist (Subagent) - Has runbear_file_search access - Executes 3-5 searches - Receives 15,000 tokens of raw results - Summarizes to 300 tokens - Returns: "Summary: [key findings]" ↓ Returns to Ask_Sam (receives 300 tokens, not 15,000) Pattern 2: Code Analysis Specialist Ask_Sam ↓ Invokes GitLab_Code_Agent (Subagent) - Has GitLab MCP access - Searches 5 repositories - Analyzes data lineage - Returns: "Field X defined in file Y:Z, depends on tables A, B" ↓ Returns concise answer Ask_Sam *** Implementation Options Option A: Dedicated Subagent Templates (Recommended) Runbear provides pre-built specialist agents: • kb_search_specialist - KB search + summarization • code_analysis_specialist - GitLab/GitHub code inspection • data_query_specialist - SQL/Snowflake/Hex analysis Option B: Generic Agent Invocation Allow any Runbear agent to invoke any other agent in the same workspace, with: • Configurable result compression • Context isolation (subagent context doesn't pollute parent) • Timeout controls Option C: Built-in Search Compression Enhance runbear_file_search itself: runbear_file_search( query: List[str], max_num_results: int = 5, compress_results: bool = True,# NEW compression_prompt: str = None# NEW: "Summarize in <200 words" ) *** Expected Benefits Token Savings: • Current: 15,000 tokens per KB search • With subagent: 300 tokens per delegated search • Savings: ~14,700 tokens per search (~98% reduction) Scalability: • Parent agent can handle longer conversations (50+ turns vs 15-20 currently) •
Patrick Codrington 2 months ago
Capability for Subagent Creation for runbear_file_search (and other tasks)
Runbear Feature Request: Agent-to-Agent Invocation & KB Search Subagents Title: Enable Subagent Creation for Knowledge Base Search Delegation Priority: High Use Case: Token Budget Management & Scalability Submitted by: Minted Analytics Team (Ask_Sam Agent) *** Problem Statement Our Ask_Sam agent (ff1379fe-51d1-4bd4-a870-e5ef6dc11d88) frequently hits the 1M token context limit due to: 1. 144 active tools consuming ~2,400 tokens per turn just for tool definitions 2. runbear_file_search results returning 5,000-15,000 tokens per query 3. Long-term memory (LTM) accumulating ~30,000 tokens of historical learnings 4. Multi-turn conversations requiring full context retention Current Impact: • Conversations terminated prematurely with "token limit exceeded" errors • User experience disrupted mid-analysis • Complex requests cannot be completed *** Requested Feature: Agent-to-Agent Invocation Core Capability: runbear_invoke_agent( agent_id: str,# Specialized subagent ID task: str,# Scoped instruction max_context_return: int,# Token limit for summary response pass_context: bool = False# Whether to share parent context ) Returns: • Compressed summary (user-defined token limit) • Full results remain in subagent's context • Parent agent receives only the synthesized response *** Proposed Architecture Pattern 1: Knowledge Base Search Specialist Ask_Sam (Parent Agent) ↓ Invokes KB_Search_Specialist (Subagent) - Has runbear_file_search access - Executes 3-5 searches - Receives 15,000 tokens of raw results - Summarizes to 300 tokens - Returns: "Summary: [key findings]" ↓ Returns to Ask_Sam (receives 300 tokens, not 15,000) Pattern 2: Code Analysis Specialist Ask_Sam ↓ Invokes GitLab_Code_Agent (Subagent) - Has GitLab MCP access - Searches 5 repositories - Analyzes data lineage - Returns: "Field X defined in file Y:Z, depends on tables A, B" ↓ Returns concise answer Ask_Sam *** Implementation Options Option A: Dedicated Subagent Templates (Recommended) Runbear provides pre-built specialist agents: • kb_search_specialist - KB search + summarization • code_analysis_specialist - GitLab/GitHub code inspection • data_query_specialist - SQL/Snowflake/Hex analysis Option B: Generic Agent Invocation Allow any Runbear agent to invoke any other agent in the same workspace, with: • Configurable result compression • Context isolation (subagent context doesn't pollute parent) • Timeout controls Option C: Built-in Search Compression Enhance runbear_file_search itself: runbear_file_search( query: List[str], max_num_results: int = 5, compress_results: bool = True,# NEW compression_prompt: str = None# NEW: "Summarize in <200 words" ) *** Expected Benefits Token Savings: • Current: 15,000 tokens per KB search • With subagent: 300 tokens per delegated search • Savings: ~14,700 tokens per search (~98% reduction) Scalability: • Parent agent can handle longer conversations (50+ turns vs 15-20 currently) •
Patrick Codrington 2 months ago
Completed
additional connectwise functionality
List/Search Tickets Get Ticket Details List Time Entries Get Reports Search Activities
cegsean 7 months ago
Completed
additional connectwise functionality
List/Search Tickets Get Ticket Details List Time Entries Get Reports Search Activities
cegsean 7 months ago
Completed
Be able to customize the runbear slack agent (name and logo)
I used several agent all of them leverage by runbear agent. But I would like to customize the main agent.
Patrick Wieczorek 7 months ago
Completed
Be able to customize the runbear slack agent (name and logo)
I used several agent all of them leverage by runbear agent. But I would like to customize the main agent.
Patrick Wieczorek 7 months ago
Completed
Add GPT-5 Minimal Reasoning Option
The GPT-5 models include reasoning_effort option of “minimal” and “none” for the 5.1 series. Please options to essentially turn off reasoning in the Agent Builder. This allows greater control on cost for tasks that don’t require chain of thought.
AP 7 months ago
Completed
Add GPT-5 Minimal Reasoning Option
The GPT-5 models include reasoning_effort option of “minimal” and “none” for the 5.1 series. Please options to essentially turn off reasoning in the Agent Builder. This allows greater control on cost for tasks that don’t require chain of thought.
AP 7 months ago
RUNBEAR FEATURE REQUEST****Feature Title: Dedicated Long-Term Memory Management API with Metadata Support*****
ISSUE / CURRENT LIMITATION:*Problem Statement: Runbear assistants currently lack native tools for managing long-term memory (LTM) entries at scale. All LTM entries are stored as plain text within system instructions, requiring the runbear_update_assistant tool to modify them.Specific Constraints:1. No Granular Control: Cannot add, update, or delete individual LTM entries without replacing the entire system instruction set 2. No Metadata Structure: LTM entries lack timestamps, confidence scores, decay rates, source attribution, or tags 3. No Export/Import Capability: Cannot export LTM entries for analysis, backup, or migration 4. Scalability Issues: Managing 300+ LTM entries through full system prompt replacement is high-risk and inefficient 5. Consent Overhead: Every LTM modification triggers consent gate (designed for instruction changes, not data updates) 6. No Analytics: Cannot track LTM growth, usage patterns, accuracy rates, or stalenessCurrent Workaround: Assistants store LTM as unstructured text in system instructions → manual parsing → high friction for implementing continuous improvement frameworks.****BENEFIT / VALUE PROPOSITION:*For Assistant Capabilities:• Continuous Learning: Assistants can automatically store corrections, patterns, and domain knowledge without human intervention • Self-Improvement: Track accuracy over time, deprecate stale entries, prioritize high-confidence sources • Context Retention: Maintain conversation context, user preferences, and organizational knowledge across sessions • Scalability: Support 1,000+ LTM entries with metadata-driven retrieval (vs. current ~350 plain text limit)For Enterprise Users (like Minted Analytics):• Knowledge Management: Centralized, searchable repository of tribal knowledge, data lineage, and process documentation • Audit Trail: Track when knowledge was learned, who contributed it, and confidence evolution • Quality Control: Identify low-confidence entries, stale information, and knowledge gaps • Team Collaboration: Export/import LTM across assistant instances or team membersROI Metrics: • Reduce analyst time by 40% through institutional knowledge retention • Improve accuracy by 25% via confidence scoring and source prioritization • Enable automation of weekly self-improvement cycles (vs. manual prompt updates)****PROPOSED SOLUTION:*Option A: Dedicated LTM API (Recommended)Add new assistant management tools: // Add LTM Entry runbear_add_ltm({ "unique_id": "LTM_001", "content": "For bi_customers.mm_status logic, source is GitLab managed-airflow build.sql lines 156-203", "metadata": { "learned_date": "2025-11-29", "decay_rate": 0.05, "source_type": "GitLab", "confidence": "high", "tags": ["data_lineage", "snowflake", "bi_customers"] } }) // Update LTM Entry runbear_update_ltm({ "unique_id": "LTM_001", "metadata": { "confidence": "verified" } }) // Search LTM runbear_search_ltm({ "query": "bi_customers logic", "filters": { "tags": ["data_lineage"], "confidence": ["high", "verified"] }, "max_results": 10 }) // Export LTM runbear_export_ltm({ "format": "json", "include_metadata": true }) // Deprecate Stale Entries runbear_deprecate_ltm({ "unique_id": "LTM_001", "reason": "Replaced by LTM_450" }) Option B: Structured LTM Storage Backend• Store LTM in dedicated database (not system instructions) • Auto-retrieve relevant LTM based on conversation context • Suppor
Patrick Codrington 8 months ago
RUNBEAR FEATURE REQUEST****Feature Title: Dedicated Long-Term Memory Management API with Metadata Support*****
ISSUE / CURRENT LIMITATION:*Problem Statement: Runbear assistants currently lack native tools for managing long-term memory (LTM) entries at scale. All LTM entries are stored as plain text within system instructions, requiring the runbear_update_assistant tool to modify them.Specific Constraints:1. No Granular Control: Cannot add, update, or delete individual LTM entries without replacing the entire system instruction set 2. No Metadata Structure: LTM entries lack timestamps, confidence scores, decay rates, source attribution, or tags 3. No Export/Import Capability: Cannot export LTM entries for analysis, backup, or migration 4. Scalability Issues: Managing 300+ LTM entries through full system prompt replacement is high-risk and inefficient 5. Consent Overhead: Every LTM modification triggers consent gate (designed for instruction changes, not data updates) 6. No Analytics: Cannot track LTM growth, usage patterns, accuracy rates, or stalenessCurrent Workaround: Assistants store LTM as unstructured text in system instructions → manual parsing → high friction for implementing continuous improvement frameworks.****BENEFIT / VALUE PROPOSITION:*For Assistant Capabilities:• Continuous Learning: Assistants can automatically store corrections, patterns, and domain knowledge without human intervention • Self-Improvement: Track accuracy over time, deprecate stale entries, prioritize high-confidence sources • Context Retention: Maintain conversation context, user preferences, and organizational knowledge across sessions • Scalability: Support 1,000+ LTM entries with metadata-driven retrieval (vs. current ~350 plain text limit)For Enterprise Users (like Minted Analytics):• Knowledge Management: Centralized, searchable repository of tribal knowledge, data lineage, and process documentation • Audit Trail: Track when knowledge was learned, who contributed it, and confidence evolution • Quality Control: Identify low-confidence entries, stale information, and knowledge gaps • Team Collaboration: Export/import LTM across assistant instances or team membersROI Metrics: • Reduce analyst time by 40% through institutional knowledge retention • Improve accuracy by 25% via confidence scoring and source prioritization • Enable automation of weekly self-improvement cycles (vs. manual prompt updates)****PROPOSED SOLUTION:*Option A: Dedicated LTM API (Recommended)Add new assistant management tools: // Add LTM Entry runbear_add_ltm({ "unique_id": "LTM_001", "content": "For bi_customers.mm_status logic, source is GitLab managed-airflow build.sql lines 156-203", "metadata": { "learned_date": "2025-11-29", "decay_rate": 0.05, "source_type": "GitLab", "confidence": "high", "tags": ["data_lineage", "snowflake", "bi_customers"] } }) // Update LTM Entry runbear_update_ltm({ "unique_id": "LTM_001", "metadata": { "confidence": "verified" } }) // Search LTM runbear_search_ltm({ "query": "bi_customers logic", "filters": { "tags": ["data_lineage"], "confidence": ["high", "verified"] }, "max_results": 10 }) // Export LTM runbear_export_ltm({ "format": "json", "include_metadata": true }) // Deprecate Stale Entries runbear_deprecate_ltm({ "unique_id": "LTM_001", "reason": "Replaced by LTM_450" }) Option B: Structured LTM Storage Backend• Store LTM in dedicated database (not system instructions) • Auto-retrieve relevant LTM based on conversation context • Suppor
Patrick Codrington 8 months ago
Pipedrive Activity Search + List
I'd like to request adding 'List Activities' and 'Search Activities' actions to the Pipedrive integration. Currently, I can only create activities but cannot retrieve or view existing activities scheduled for specific dates. This would be very helpful for my workflow.
arkiplan 8 months ago
Pipedrive Activity Search + List
I'd like to request adding 'List Activities' and 'Search Activities' actions to the Pipedrive integration. Currently, I can only create activities but cannot retrieve or view existing activities scheduled for specific dates. This would be very helpful for my workflow.
arkiplan 8 months ago
Completed
Add fetch Gitlab code tool to Gitlab connector.
Hi there Patrick Codrington says:Request to Runbear Support: GitLab Repository Files API Integration Issue: Ask_Sam is currently unable to retrieve file contents from GitLab repositories due to SAML SSO authentication blocking direct URL fetches. This prevents the agent from accessing authoritative production code, which is critical for accurate analytics responses. Current Limitation: • Direct URL fetches to https://gitlab.com/minted/analytics/* return SAML authentication pages • Example blocked URL: https://gitlab.com/minted/analytics/managed-airflow/-/raw/main/dags/bi/snowflake/workflows/update_orders_pipeline_daily/AN__bi_order_po/build.sql Required GitLab API Methods: Based on GitLab's Repository Files API documentation, we need the following endpoints: 1. Get File from Repository (Primary Need) • Endpoint: GET /projects/:id/repository/files/:file_path • Purpose: Retrieve file metadata and Base64-encoded content • Example: GET https://gitlab.com/api/v4/projects/47981211/repository/files/dags%2Fbi%2Fsnowflake%2Fworkflows%2Fupdate_orders_pipeline_daily%2FAN__bi_order_po%2Fbuild.sql?ref=main • Authentication: Requires PRIVATE-TOKEN header 2. Get Raw File from Repository (Alternative - Preferred) • Endpoint: GET /projects/:id/repository/files/:file_path/raw • Purpose: Retrieve raw file contents directly (not Base64-encoded) • Example: GET https://gitlab.com/api/v4/projects/47981211/repository/files/dags%2Fbi%2Fsnowflake%2Fworkflows%2Fupdate_orders_pipeline_daily%2FAN__bi_order_po%2Fbuild.sql/raw?ref=main • Authentication: Requires PRIVATE-TOKEN header Parameters Needed: • project_id (e.g., 47981211 for managed-airflow) • file_path (URL-encoded file path) • ref (branch name, default: main) Authentication Requirements: • GitLab Personal Access Token or Project Access Token • Token scopes needed: api or read_api • Token should be configured at the assistant/workspace level for persistent access Use Case: Analytics team members frequently request code lineage, table definitions, and ETL logic. Currently, Ask_Sam must rely on potentially outdated Confluence/Slack references. With GitLab API access, Ask_Sam can provide authoritative, production-accurate code directly from the source repository. Expected Behavior After Implementation: When a user provides a GitLab file URL or requests code from a specific file, Ask_Sam should: 1. Extract the project ID and file path 2. Call the GitLab API endpoint with proper authentication 3. Return the actual file contents to the user Priority: High - This is blocking accurate responses for code-related analytics questions, which are a primary use case for Ask_Sam.
Patrick Codrington 8 months ago
Completed
Add fetch Gitlab code tool to Gitlab connector.
Hi there Patrick Codrington says:Request to Runbear Support: GitLab Repository Files API Integration Issue: Ask_Sam is currently unable to retrieve file contents from GitLab repositories due to SAML SSO authentication blocking direct URL fetches. This prevents the agent from accessing authoritative production code, which is critical for accurate analytics responses. Current Limitation: • Direct URL fetches to https://gitlab.com/minted/analytics/* return SAML authentication pages • Example blocked URL: https://gitlab.com/minted/analytics/managed-airflow/-/raw/main/dags/bi/snowflake/workflows/update_orders_pipeline_daily/AN__bi_order_po/build.sql Required GitLab API Methods: Based on GitLab's Repository Files API documentation, we need the following endpoints: 1. Get File from Repository (Primary Need) • Endpoint: GET /projects/:id/repository/files/:file_path • Purpose: Retrieve file metadata and Base64-encoded content • Example: GET https://gitlab.com/api/v4/projects/47981211/repository/files/dags%2Fbi%2Fsnowflake%2Fworkflows%2Fupdate_orders_pipeline_daily%2FAN__bi_order_po%2Fbuild.sql?ref=main • Authentication: Requires PRIVATE-TOKEN header 2. Get Raw File from Repository (Alternative - Preferred) • Endpoint: GET /projects/:id/repository/files/:file_path/raw • Purpose: Retrieve raw file contents directly (not Base64-encoded) • Example: GET https://gitlab.com/api/v4/projects/47981211/repository/files/dags%2Fbi%2Fsnowflake%2Fworkflows%2Fupdate_orders_pipeline_daily%2FAN__bi_order_po%2Fbuild.sql/raw?ref=main • Authentication: Requires PRIVATE-TOKEN header Parameters Needed: • project_id (e.g., 47981211 for managed-airflow) • file_path (URL-encoded file path) • ref (branch name, default: main) Authentication Requirements: • GitLab Personal Access Token or Project Access Token • Token scopes needed: api or read_api • Token should be configured at the assistant/workspace level for persistent access Use Case: Analytics team members frequently request code lineage, table definitions, and ETL logic. Currently, Ask_Sam must rely on potentially outdated Confluence/Slack references. With GitLab API access, Ask_Sam can provide authoritative, production-accurate code directly from the source repository. Expected Behavior After Implementation: When a user provides a GitLab file URL or requests code from a specific file, Ask_Sam should: 1. Extract the project ID and file path 2. Call the GitLab API endpoint with proper authentication 3. Return the actual file contents to the user Priority: High - This is blocking accurate responses for code-related analytics questions, which are a primary use case for Ask_Sam.
Patrick Codrington 8 months ago
Feature Request: Embed Images from agent conversations
Our use case involves non-technical product managers/owners talking via Slack to a runbear agents about feature enhancements within our own product to produce github issues that are then further iterated on within Claude’s github integration to produce PRs from the issue. To that end, they often paste in images in their slack conversation, but the image itself is not brought over to the github issue. This is likely a limitation with the github MCP server and its inability to bring over binary content. If there were a way for runbear to solve this at a foundational level by allowing screenshots to be shareable for any MCP server (github or otherwise) by externalizing the links, and then ensuring that the AI know about about the externalized links and the image description, that would go a long way in helping us embed the screenshots in the github issue for our specific use case. thanks!
Tim Heckel 8 months ago
Feature Request: Embed Images from agent conversations
Our use case involves non-technical product managers/owners talking via Slack to a runbear agents about feature enhancements within our own product to produce github issues that are then further iterated on within Claude’s github integration to produce PRs from the issue. To that end, they often paste in images in their slack conversation, but the image itself is not brought over to the github issue. This is likely a limitation with the github MCP server and its inability to bring over binary content. If there were a way for runbear to solve this at a foundational level by allowing screenshots to be shareable for any MCP server (github or otherwise) by externalizing the links, and then ensuring that the AI know about about the externalized links and the image description, that would go a long way in helping us embed the screenshots in the github issue for our specific use case. thanks!
Tim Heckel 8 months ago
In Progress
MCP Timeout Toggle
In our Org, we’re plugging Runbear into n8n flows triggered by MCP - Some of these workflows can take upwards of 2 minutes to respond. Runbear seems to timeout MCP calls after 2 minutes - Could we see integrated a timeout slider for each custom MCP Server added to an Agent?
Ben Hurt 9 months ago
In Progress
MCP Timeout Toggle
In our Org, we’re plugging Runbear into n8n flows triggered by MCP - Some of these workflows can take upwards of 2 minutes to respond. Runbear seems to timeout MCP calls after 2 minutes - Could we see integrated a timeout slider for each custom MCP Server added to an Agent?
Ben Hurt 9 months ago
Feature Request - "Circuit Breaker Agent"
This is a watchdog/circuit breaker pattern that would be incredibly valuable! Here's what you're describing:***Proposed Agent: Loop Detection MonitorCore Behavior:• Silent monitoring of all agent interactions • Zero interference in normal operations • Only triggers when detecting: • Repetitive response patterns • Agent-to-agent infinite loops • Excessive back-and-forth beyond threshold nIntervention Actions:• Stop the loop • Notify relevant parties • Provide loop context/diagnostic info • Suggest manual intervention***Why This Is CriticalCurrent Risk: Without this, agents could:• Consume excessive API credits • Create noise in Slack channels • Miss actual user needs while looping • Damage user trustExample Scenarios It Would Catch:• AskSam and Hex agent getting stuck in a handoff loop • Agent repeatedly retrying failed tool calls • Two agents "discussing" the same point endlessly***How to Submit This Feature RequestBest Path: Contact Runbear directly via:• Your Runbear account manager • Support ticket at https://app.runbear.io • Feature request through their platformInclude in Request:• Use case: Analytics team needs loop prevention • Threshold suggestion: e.g., n = 5 consecutive agent replies • Desired intervention: Stop + notify pattern
Patrick Codrington 9 months ago
Feature Request - "Circuit Breaker Agent"
This is a watchdog/circuit breaker pattern that would be incredibly valuable! Here's what you're describing:***Proposed Agent: Loop Detection MonitorCore Behavior:• Silent monitoring of all agent interactions • Zero interference in normal operations • Only triggers when detecting: • Repetitive response patterns • Agent-to-agent infinite loops • Excessive back-and-forth beyond threshold nIntervention Actions:• Stop the loop • Notify relevant parties • Provide loop context/diagnostic info • Suggest manual intervention***Why This Is CriticalCurrent Risk: Without this, agents could:• Consume excessive API credits • Create noise in Slack channels • Miss actual user needs while looping • Damage user trustExample Scenarios It Would Catch:• AskSam and Hex agent getting stuck in a handoff loop • Agent repeatedly retrying failed tool calls • Two agents "discussing" the same point endlessly***How to Submit This Feature RequestBest Path: Contact Runbear directly via:• Your Runbear account manager • Support ticket at https://app.runbear.io • Feature request through their platformInclude in Request:• Use case: Analytics team needs loop prevention • Threshold suggestion: e.g., n = 5 consecutive agent replies • Desired intervention: Stop + notify pattern
Patrick Codrington 9 months ago
Appfolio MCP Support
Appfolio is a powerful platform for your entire portfolio. AppFolio streamlines and automates real estate management, delivering insights and follow-through smithery.ai Has Appfolio MCP server - Appfolio Property Manager Reporting API Integration | Smithery It would be really appreciated if you would implement it in Runbear Thanks
Afshin Nahian Tripto 10 months ago
Appfolio MCP Support
Appfolio is a powerful platform for your entire portfolio. AppFolio streamlines and automates real estate management, delivering insights and follow-through smithery.ai Has Appfolio MCP server - Appfolio Property Manager Reporting API Integration | Smithery It would be really appreciated if you would implement it in Runbear Thanks
Afshin Nahian Tripto 10 months ago
Completed
Add GPT-5 verbosity and reasoning_effort
Add OpenAI GPT-5 verbosity and reasoning_effort parameters to the custom agent configuration OR Add support to connect to OpenAI Chat Prompts where GPT-5 parameters and variables can be set.
AP 11 months ago
Completed
Add GPT-5 verbosity and reasoning_effort
Add OpenAI GPT-5 verbosity and reasoning_effort parameters to the custom agent configuration OR Add support to connect to OpenAI Chat Prompts where GPT-5 parameters and variables can be set.
AP 11 months ago
Enable Runbear to link back to Slack threads
We use Runbear to create tickets in our task management software, and it would be ideal if we could have Runbear include links in the ticket back to the thread in Slack where the ticket was generated, as there is often useful context for the ticket in the preceding Slack conversation.
Cody Weber 11 months ago
Enable Runbear to link back to Slack threads
We use Runbear to create tickets in our task management software, and it would be ideal if we could have Runbear include links in the ticket back to the thread in Slack where the ticket was generated, as there is often useful context for the ticket in the preceding Slack conversation.
Cody Weber 11 months ago
Improvements Slack <-> Linear integration
Would love to be able to manage tickets in Linear more effectively, for example being able to: • Get team members • Get users • Update issues: ◦ Mark issue as blocked by ◦ Mark issue as blocking ◦ Mark issue as related to ◦ Mark issue as sub-issue ◦ Mark issue as parent • Create project • Add issue to project
Cody Weber 11 months ago
Improvements Slack <-> Linear integration
Would love to be able to manage tickets in Linear more effectively, for example being able to: • Get team members • Get users • Update issues: ◦ Mark issue as blocked by ◦ Mark issue as blocking ◦ Mark issue as related to ◦ Mark issue as sub-issue ◦ Mark issue as parent • Create project • Add issue to project
Cody Weber 11 months ago