Agent Tools Reference
This skill covers the tools available to Overcut agents, built-in agent type presets, and the persistent memory system.
Tool Categories Overview
Overcut provides 40 user-configurable tools plus 10 system tools, organized by category:
| Category |
Tools |
Count |
| File System |
CreateDirectory, ListDirectory, ReadFile, WriteFile, DeleteFile, EditFile, AppendFile |
7 |
| Code Utilities |
ListCodeDefinitionNames, CodeSearch, SemanticCodeSearch, RunTerminalCmd, GetDiagnostics |
5 |
| Tickets/Issues |
CreateTicket, UpdateTicket, ListTickets, ReadTicket, AddCommentToTicket, UpdateCommentOnTicket |
6 |
| Pull Requests |
CreatePullRequest, ReadPullRequest, UpdatePullRequest, ListPullRequests, MergePullRequest, AddCommentToPullRequest, UpdateCommentOnPullRequest, ClosePullRequest |
8 |
| Code Review |
AddPullRequestReviewThread, SubmitReview, AddPullRequestReviewThreadReply, GetPullRequestDiff, GetPullRequestDiffLineNumbers |
5 |
| CI/CD |
GetCiRunDetails, GetCiRunLogs, ListPrCiRuns, GetCiJobLogs, RetryCiWorkflow |
5 |
| Attachments |
GetTicketAttachments, GetPullRequestAttachments |
2 |
| Exploration |
ExploreCodebase |
1 |
| Memory (auto-injected) |
MemoryWrite, MemoryRead |
2 |
| Scratchpad (auto-injected) |
WriteScratchpad, ReadScratchpad, ListScratchpads, AppendScratchpad |
4 |
| Coordinator-only (auto-injected) |
DelegateToSubAgent, UpdateStatus, TaskCompletedTool |
3 |
Built-in Agent Types
When creating agents in Overcut, you select a base type that determines the default tool set:
| Agent Type |
Tools |
Focus |
| SeniorDeveloper |
31 tools |
Full-stack development: all file, code, ticket, PR, CI/CD, and exploration tools |
| CodeReview |
26 tools |
Code review: read-only file access + append, code search, review tools, PR read/update, CI read tools |
| TechWriter |
25 tools |
Documentation: full file access, code search, ticket and PR tools |
| ProductManager |
20 tools |
Planning: read-only file access, code search, ticket CRUD, PR read/update |
| ExploreAgent |
5 tools |
Exploration: ReadFile, ListDirectory, CodeSearch, SemanticCodeSearch, RunTerminalCmd |
Key differences:
- SeniorDeveloper can create PRs, merge PRs, delete files — full write access
- CodeReview cannot create PRs or delete files, but can submit reviews and append to files
- TechWriter has nearly the same tools as SeniorDeveloper (full write access for docs)
- ProductManager cannot create PRs, write files, or delete files — primarily read + ticket management
- ExploreAgent is minimal — read-only code exploration only
Automatic Tool Injection
Regardless of base type, these tools are automatically added at runtime:
- Memory tools (
memory_write, memory_read) — added to all agents (coordinator and sub-agents)
- Scratchpad tools (
write_scratchpad, read_scratchpad, list_scratchpads, append_scratchpad) — added to all agents (coordinator and sub-agents)
- Coordinator tools (
delegate_to_sub_agent, update_status, task_completed) — added only to coordinator agents in agent.session steps
Memory System
The memory system allows agents to persist and retrieve learnings across workflow executions.
memory_write
{
"title": "Short summary of the learning (1 line)",
"content": "Full details of the learning"
}
title: Short summary used as the identifier (1 line)
content: Full details of the learning
memory_read
{
"title": "Memory title to retrieve"
}
title: Case-insensitive match against stored memory titles
Memory Behavior
- Memories persist across workflow executions for the same agent
- Auto-injected into agent context at the start of each execution
- Use for: repository conventions, past review patterns, common issues, team preferences
- Keep titles descriptive and unique for reliable retrieval
Scratchpad System
The scratchpad system provides ephemeral, per-run storage for sharing structured data between agents within a single workflow execution.
write_scratchpad
{
"name": "Scratchpad name (e.g., 'review-findings', 'chunk-1')",
"content": "Content to write (overwrites existing)"
}
name: Identifier for the scratchpad (use kebab-case, descriptive names)
content: Full content to write (replaces any existing content)
read_scratchpad
{
"name": "Scratchpad name to read"
}
name: Exact name of the scratchpad to read
list_scratchpads
{}
No parameters. Returns all scratchpad names in the current run.
append_scratchpad
{
"name": "Scratchpad name to append to",
"content": "Content to append"
}
name: Scratchpad to append to (creates it if it doesn't exist)
content: Content to append (added after existing content)
Scratchpad Behavior
- Scratchpads are ephemeral — they exist only for the duration of the workflow run
- Auto-injected into all agents (coordinator and sub-agents)
- Use for: intermediate findings, chunk data, structured results shared between steps or agents
append_scratchpad is safe for concurrent writes from parallel agents
- Use kebab-case names (e.g.,
review-findings, chunk-1, security-analysis)
Scratchpad vs Memory
|
Scratchpad |
Memory |
| Lifetime |
Single workflow run |
Persists across runs |
| Purpose |
Share data between agents/steps |
Store learnings for future runs |
| Concurrency |
append_scratchpad is safe for parallel writes |
Not designed for concurrent access |
| Use when |
Passing intermediate data (findings, chunks) |
Saving patterns, conventions, preferences |
Tool Selection for Prompts
When writing step prompts, restrict tools to only what's needed. Use allowed/prohibited tool tables:
### Allowed Tools
| Tool | Purpose |
|------|---------|
| `read_file` | Read source files for analysis |
| `code_search` | Find relevant code patterns |
| `append_scratchpad` | Store findings for downstream steps |
| `add_comment_to_pull_request` | Post review summary |
### Prohibited Tools
❌ `write_file` — this is a read-only analysis step
❌ `run_terminal_cmd` — no terminal commands needed
❌ `create_pull_request` — PR already exists
This pattern appears in many real playbooks and significantly improves agent focus and efficiency.
For the complete tool listing with tool identifiers, see references/tool-catalog.md.
Converted and distributed by TomeVault — claim your Tome and manage your conversions.
1---2name: overcut-ai-overcut-playbooks-agent-tools-reference3description: Agent Tools Reference4---56# Agent Tools Reference78This skill covers the tools available to Overcut agents, built-in agent type presets, and the persistent memory system.910## Tool Categories Overview1112Overcut provides 40 user-configurable tools plus 10 system tools, organized by category:1314| Category | Tools | Count |15|----------|-------|-------|16| **File System** | CreateDirectory, ListDirectory, ReadFile, WriteFile, DeleteFile, EditFile, AppendFile | 7 |17| **Code Utilities** | ListCodeDefinitionNames, CodeSearch, SemanticCodeSearch, RunTerminalCmd, GetDiagnostics | 5 |18| **Tickets/Issues** | CreateTicket, UpdateTicket, ListTickets, ReadTicket, AddCommentToTicket, UpdateCommentOnTicket | 6 |19| **Pull Requests** | CreatePullRequest, ReadPullRequest, UpdatePullRequest, ListPullRequests, MergePullRequest, AddCommentToPullRequest, UpdateCommentOnPullRequest, ClosePullRequest | 8 |20| **Code Review** | AddPullRequestReviewThread, SubmitReview, AddPullRequestReviewThreadReply, GetPullRequestDiff, GetPullRequestDiffLineNumbers | 5 |21| **CI/CD** | GetCiRunDetails, GetCiRunLogs, ListPrCiRuns, GetCiJobLogs, RetryCiWorkflow | 5 |22| **Attachments** | GetTicketAttachments, GetPullRequestAttachments | 2 |23| **Exploration** | ExploreCodebase | 1 |24| **Memory** (auto-injected) | MemoryWrite, MemoryRead | 2 |25| **Scratchpad** (auto-injected) | WriteScratchpad, ReadScratchpad, ListScratchpads, AppendScratchpad | 4 |26| **Coordinator-only** (auto-injected) | DelegateToSubAgent, UpdateStatus, TaskCompletedTool | 3 |2728## Built-in Agent Types2930When creating agents in Overcut, you select a base type that determines the default tool set:3132| Agent Type | Tools | Focus |33|------------|-------|-------|34| **SeniorDeveloper** | 31 tools | Full-stack development: all file, code, ticket, PR, CI/CD, and exploration tools |35| **CodeReview** | 26 tools | Code review: read-only file access + append, code search, review tools, PR read/update, CI read tools |36| **TechWriter** | 25 tools | Documentation: full file access, code search, ticket and PR tools |37| **ProductManager** | 20 tools | Planning: read-only file access, code search, ticket CRUD, PR read/update |38| **ExploreAgent** | 5 tools | Exploration: ReadFile, ListDirectory, CodeSearch, SemanticCodeSearch, RunTerminalCmd |3940**Key differences:**41- **SeniorDeveloper** can create PRs, merge PRs, delete files — full write access42- **CodeReview** cannot create PRs or delete files, but can submit reviews and append to files43- **TechWriter** has nearly the same tools as SeniorDeveloper (full write access for docs)44- **ProductManager** cannot create PRs, write files, or delete files — primarily read + ticket management45- **ExploreAgent** is minimal — read-only code exploration only4647### Automatic Tool Injection4849Regardless of base type, these tools are **automatically added** at runtime:50- **Memory tools** (`memory_write`, `memory_read`) — added to all agents (coordinator and sub-agents)51- **Scratchpad tools** (`write_scratchpad`, `read_scratchpad`, `list_scratchpads`, `append_scratchpad`) — added to all agents (coordinator and sub-agents)52- **Coordinator tools** (`delegate_to_sub_agent`, `update_status`, `task_completed`) — added only to coordinator agents in `agent.session` steps5354## Memory System5556The memory system allows agents to persist and retrieve learnings across workflow executions.5758### memory_write5960```json61{62 "title": "Short summary of the learning (1 line)",63 "content": "Full details of the learning"64}65```6667- `title`: Short summary used as the identifier (1 line)68- `content`: Full details of the learning6970### memory_read7172```json73{74 "title": "Memory title to retrieve"75}76```7778- `title`: Case-insensitive match against stored memory titles7980### Memory Behavior8182- Memories persist across workflow executions for the same agent83- Auto-injected into agent context at the start of each execution84- Use for: repository conventions, past review patterns, common issues, team preferences85- Keep titles descriptive and unique for reliable retrieval8687## Scratchpad System8889The scratchpad system provides ephemeral, per-run storage for sharing structured data between agents within a single workflow execution.9091### write_scratchpad9293```json94{95 "name": "Scratchpad name (e.g., 'review-findings', 'chunk-1')",96 "content": "Content to write (overwrites existing)"97}98```99100- `name`: Identifier for the scratchpad (use kebab-case, descriptive names)101- `content`: Full content to write (replaces any existing content)102103### read_scratchpad104105```json106{107 "name": "Scratchpad name to read"108}109```110111- `name`: Exact name of the scratchpad to read112113### list_scratchpads114115```json116{}117```118119No parameters. Returns all scratchpad names in the current run.120121### append_scratchpad122123```json124{125 "name": "Scratchpad name to append to",126 "content": "Content to append"127}128```129130- `name`: Scratchpad to append to (creates it if it doesn't exist)131- `content`: Content to append (added after existing content)132133### Scratchpad Behavior134135- Scratchpads are **ephemeral** — they exist only for the duration of the workflow run136- Auto-injected into all agents (coordinator and sub-agents)137- Use for: intermediate findings, chunk data, structured results shared between steps or agents138- `append_scratchpad` is safe for concurrent writes from parallel agents139- Use kebab-case names (e.g., `review-findings`, `chunk-1`, `security-analysis`)140141### Scratchpad vs Memory142143| | Scratchpad | Memory |144|---|---|---|145| **Lifetime** | Single workflow run | Persists across runs |146| **Purpose** | Share data between agents/steps | Store learnings for future runs |147| **Concurrency** | `append_scratchpad` is safe for parallel writes | Not designed for concurrent access |148| **Use when** | Passing intermediate data (findings, chunks) | Saving patterns, conventions, preferences |149150## Tool Selection for Prompts151152When writing step prompts, restrict tools to only what's needed. Use allowed/prohibited tool tables:153154```markdown155### Allowed Tools156| Tool | Purpose |157|------|---------|158| `read_file` | Read source files for analysis |159| `code_search` | Find relevant code patterns |160| `append_scratchpad` | Store findings for downstream steps |161| `add_comment_to_pull_request` | Post review summary |162163### Prohibited Tools164❌ `write_file` — this is a read-only analysis step165❌ `run_terminal_cmd` — no terminal commands needed166❌ `create_pull_request` — PR already exists167```168169This pattern appears in many real playbooks and significantly improves agent focus and efficiency.170171For the complete tool listing with tool identifiers, see `references/tool-catalog.md`.172173---174> Converted and distributed by [TomeVault](https://tomevault.io/claim/overcut-ai) — claim your Tome and manage your conversions.175<!-- tomevault:4.0:skill_md:2026-04-11 -->