Automated summary
Manages ISCP dispatches — the inter-agent messaging inbox indexed in a SQLite database — through a single shell command to list, read, reply to, create, and resolve messages.
Strengths
- Covers the full message lifecycle: list, fetch, read, reply, create, resolve, status
- Clearly separates `fetch` (peek without status change) from `read` (commit to processing, marks as read)
- Resolves payloads transparently across branches and worktrees, so no merge to main is needed just to read a dispatch
- `reply` and `resolve --response` handle recipient, "Re:" prefix, and in_reply_to linkage automatically
Limitations
- Entirely dependent on the project's `agency/tools/dispatch` script and ISCP database; not portable
- Assumes the repo/principal/agent address scheme specific to this infrastructure
- No error or conflict handling described for invalid IDs or already-resolved dispatches
Use it when working inside an ISCP-enabled repository and you need to check, act on, or send messages to other agents.
Avoid it on projects without the `agency/tools/dispatch` tooling or ISCP database, where the command will fail.
Security analysis
CautionThe skill is a wrapper that shells out to a local project helper for SQLite/git-backed dispatch messaging. There is no exfiltration, destructive command, obfuscated payload, or secret-sending behavior; all actions are legitimate inter-agent messaging. The risk is limited to broad inherited Bash(*) permissions, reliance on an unverified project script, and unquoted argument interpolation, which warrants caution rather than a risky rating.
- •No allowed-tools declared, so the skill inherits unrestricted Bash(*) per its own comment; this is a broad permission surface for a dispatch-management helper.
- •Wraps a project-local binary/script at $CLAUDE_PROJECT_DIR/agency/tools/dispatch whose behavior is not auditable from this file; trust depends entirely on repo contents.
- •Uses unquoted `$ARGUMENTS` interpolation into shell commands (e.g. `bash .../dispatch list $ARGUMENTS`), which is a mild shell-metacharacter injection surface, though arguments originate from the invoking user.
- •Comment actively argues for removing skill-level tool restrictions, which normalizes weaker sandboxing.
Examples
List my unread dispatches, then fetch the two most recent ones and summarize what they're asking for.Reply to dispatch 42 saying the review comments are addressed, then resolve it.Create a high-priority dispatch to acme/core/planner with subject "Schema migration blocked" and a short body explaining the failure.description: Manage dispatches — list, read, fetch, reply, create, resolve
<!-- Flag #62/#63: allowed-tools removed. Inherits Bash(*) from .claude/settings.json. Restricting to specific subcommand patterns at the skill level silently blocks agents on permission prompts the agent cannot see — see dispatch #171 for the devex incident that surfaced this trap. -->Dispatch Management
Manage ISCP dispatches for inter-agent communication. Dispatches are indexed in the ISCP SQLite DB with payload files in git. All operations use integer dispatch IDs, not file paths.
Arguments
$ARGUMENTS: Subcommand and arguments. One of:list— list dispatches addressed to current agentlist --all— list all dispatches across agentslist --status <s>— filter by status (unread, read, resolved)list --type <t>— filter by typeread <id>— display dispatch payload and mark as readfetch <id>— read-only peek (no status change)reply <id> "message"— quick response to a dispatchcreate --to <addr> --subject <text>— create a new dispatchcreate --to <addr> --subject <text> --body <text>— create with inline contentresolve <id>— mark a dispatch as resolvedcheck— silent check for unread items (hook use)status <id>— show full dispatch record
If empty, defaults to list.
Instructions
list
Run bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch list $ARGUMENTS and display results.
read / fetch
read <id>— displays the dispatch and marks it as read. Use when you're committing to process it.fetch <id>— displays the dispatch without changing status. Use to peek before deciding.
Run bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch read <id> or fetch <id>.
Summarize the content for the user.
reply
Quick response to an existing dispatch. Auto-resolves the recipient from the original sender, prefixes subject with "Re:", and sets the in_reply_to FK.
bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch reply <id> "Your response message"
create
Create a new dispatch. The --to address must be fully qualified (repo/principal/agent).
bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch create --to <addr> --subject <text> [--body <text>] [--type <type>] [--priority <p>]
- Without
--body: writes a template payload (warns about placeholders — edit before committing) - With
--body: writes the message as payload content (no template) - Types: directive, seed, review, review-response, commit, master-updated, escalation, dispatch
- Priority: normal (default), high, low
resolve
Mark a dispatch as resolved, optionally linking a response dispatch:
bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch resolve <id> [--response <id>]
check
Silent check for unread items. Returns JSON {"systemMessage": "..."} when items are waiting, silent exit 0 when empty. Used by the iscp-check hook.
status
Show the full record for a dispatch including timestamps, read_by, and payload path:
bash $CLAUDE_PROJECT_DIR/agency/tools/dispatch status <id>
Dispatch Workflow
- Receive:
dispatch list→ see what's waiting - Peek:
dispatch fetch <id>→ inspect without committing - Read:
dispatch read <id>→ commit to processing, marks as read - Reply:
dispatch reply <id> "message"→ quick acknowledgment or response - Resolve:
dispatch resolve <id>→ mark as done
Payload Resolution
Dispatch payloads are resolved transparently across branches and worktrees. The resolution ladder: local worktree → main checkout → git show all branches → scan all worktree directories. You don't need to merge main to read a dispatch — just dispatch read <id>.
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/dispatch-management-mryiqgtq/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/dispatch-management-mryiqgtq' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/dispatch-management-mryiqgtq/download' --output '.skill-downloads/dispatch-management-mryiqgtq/SKILL.md.part' && mv '.skill-downloads/dispatch-management-mryiqgtq/SKILL.md.part' '.skill-downloads/dispatch-management-mryiqgtq/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
This file does not provide a usable name/description header for this native procedure. It remains available to download and inspect; use the manual method below.
Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/dispatch-management-mryiqgtq/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
PDF Reader
Productivity
Read and extract text from PDF files, including documents, reports, contracts, and spreadsheets. Works with local files, URLs, and WhatsApp attachments. Use whenever PDF content is needed, even if not explicitly requested.
Discord Messaging Workflows
Productivity
Use the message tool to send, edit, and moderate Discord messages with interactive components.
Interactive Verdict Preview
Productivity
Generate an interactive HTML preview of keep/drop verdicts for maintainer review, then process the pasted feedback JSON.