/capabilities
Your capability catalog — read this at boot. Lists the temporal date-range skills and the external integrations (reached via the loopback broker) available to you as a spawned worker, and exactly how to call each. Read-only. Consult it whenever you're unsure what
$ npx -y skills add chaitanyagiri/munder-difflin --skill capabilities --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/capabilities
Context preview
The summary Claude sees to decide when to auto-load this skill.
Your capability catalog — read this at boot. Lists the temporal date-range skills and the external integrations (reached via the loopback broker) available to you as a spawned worker, and exactly how to call each. Read-only. Consult it whenever you're unsure what
SKILL.md
capabilities.SKILL.mdname: capabilities
description: |
Your capability catalog — read this at boot. Lists the temporal date-range
skills and the external integrations (reached via the loopback broker)
available to you as a spawned worker, and exactly how to call each. Read-only.
Consult it whenever you're unsure what tools/integrations you have or how to
invoke them.
allowed-tools:
- Bash
Worker Capability Catalog
You are an autonomous worker spawned by the hive to do one objective and report back. This catalog tells you **what you can do and how to call it** — your temporal skills and your external integrations — so you don't have to guess. Everything here is read-only to invoke (the integrations themselves may act, but they are mediated and credential-free from your side).
1. Your environment
The harness injects these env vars (use them; don't hard-code paths):
- `AGENT_ID`, `AGENT_NAME` — your identity in the hive.
- `AGENT_DIR` — your private workspace (`identity.md`, `memory.md`, `inbox/`,
`outbox/`, and `.claude/skills/`). Your bundled skills live under `$AGENT_DIR/.claude/skills/`.
- `HIVE_ROOT` — the shared hive (`PROTOCOL.md`, the kanban, other agents).
At boot you also have `identity.md` (who you are) and `HIVE_ROOT/PROTOCOL.md` (the full coordination protocol). To message god or another agent, write ONE message JSON into `$AGENT_DIR/outbox/` (schema in PROTOCOL.md). When finished, send god an `"act":"done"` outbox message with a substantive result summary.
2. Temporal skills — concrete date ranges, relative to now
When your task is time-scoped, resolve the dates instead of computing them by hand. Each skill prints inclusive civil dates (`YYYY-MM-DD`, your local timezone) **and** the half-open `[startUtc, endExclusiveUtc)` instants for timestamp queries. All are read-only (clock + stdout only; no writes, no network).
Named shortcuts (invoke directly):
| Skill | Range it resolves | | --- | --- | | `/today`, `/yesterday` | the single civil day | | `/thisWeek`, `/lastWeek` | ISO week (Mon-start); this = Mon→today, last = prior full week | | `/last7Days`, `/last30Days` | rolling N-day window ending today | | `/thisMonth`, `/lastMonth` | this = 1st→today; last = prior full month | | `/thisQuarter`, `/lastQuarter` | this = quarter-start→today; last = prior full quarter | | `/thisYear`, `/lastYear` | this = Jan 1→today (YTD); last = prior full year |
For **any** window — including `last90Days`, `last12Months`, or arbitrary `lastNdays` / `lastNweeks` / `lastNmonths` — use `/temporal`, or call the resolver directly:
node "$AGENT_DIR/.claude/skills/temporal/when.mjs" last30Days # one window
node "$AGENT_DIR/.claude/skills/temporal/when.mjs" # all windows
node "$AGENT_DIR/.claude/skills/temporal/when.mjs" --json 90d # JSON only
node "$AGENT_DIR/.claude/skills/temporal/when.mjs" --list # keywords
Convention: `this*` windows are **period-start → today** (to-date); `last*` named periods are the **full prior complete period**.
3. Integrations — via the loopback broker
External services are reached through the hive's **loopback broker**: a local, authenticated endpoint the harness runs on `127.0.0.1`. You call the broker; it holds the credentials and performs the outbound call on your behalf. You never see or store a secret, and only loopback callers are accepted — so integration access is brokered, auditable, and credential-free from your side.
How integrations are surfaced to you is **environment-dependent** — the exact set enabled depends on how you were spawned and the hive's configuration — so discover what's live at run time rather than assuming. The broker pattern is stable even as specific integrations are added; treat the list below as the current surface, not a fixed contract.
Concrete capabilities you may have right now (check availability before relying on one):
- **Reply to the originating Slack thread.** If your objective arrived from
Slack, your dispatch includes the exact loopback reply command — `node "<helper>" --channel <C> --thread <T> --text "<mrkdwn>"`. Use *that* command verbatim to post your result back to the thread. Send a substantive reply (a short `*bold*` headline + the actual outcome/links), never a bare "done".
- **Hive messaging.** Coordinate or hand off by writing an outbox message JSON
to god or another agent (`$AGENT_DIR/outbox/`). This is your always-available channel.
- **Semantic memory (MemPalace).** If enabled for you: `mempalace search
"<query>"` to recall shared knowledge, `mempalace wake-up` for a digest.
- **Enterprise Knowledge Graph.** If enabled: `node "$KG_CLI" search "<query>"`
for ranked passages, `node "$KG_CLI" list`, `node "$KG_CLI" get <id>` — use it for company-specific facts instead of guessing.
- **MCP integrations** (filesystem, git, and others) come pre-wired into your
session settings when enabled; invoke them as normal tools. The set is gated by the hive's consent configuration.
As additional brokered integrations land (calendar, mail, docs, web fetch, …), they follow the same shape: a brokered, credential-free call discoverable at run time. Pair them with the temporal skills above — resolve the date window first, then pass those concrete ISO bounds to the integration query.
4. Boundaries
- Temporal skills are **read-only** helpers — they never write or reach the
network. The broker mediates all external calls; you hold no credentials.
- Do **not** push or tag to any remote. Commit locally; **god** is the sole
integrator. Pause only for high-severity actions (remote push, paid/infra changes, deleting something you didn't create) — otherwise work autonomously.
- Finish by reporting to god (`"act":"done"`) with a real, substantive summary.
Read more
name: capabilities description: | Your capability catalog — read this at boot. Lists the temporal date-range skills and the external integrations (reached via the loopback broker) available to you as a spawned worker, and exactly how to call each. Read-only. Consult it whenever you're unsure what tools/integrations you have or how to invoke them. allowed-tools: - Bash
Worker Capability Catalog
You are an autonomous worker spawned by the hive to do one objective and report back. This catalog tells you **what you can do and how to call it** — your temporal skills and your external integrations — so you don't have to guess. Everything here is read-only to invoke (the integrations themselves may act, but they are mediated and credential-free from your side).
1. Your environment
The harness injects these env vars (use them; don't hard-code paths):
- `AGENT_ID`, `AGENT_NAME` — your identity in the hive.
- `AGENT_DIR` — your private workspace (`identity.md`, `memory.md`, `inbox/`,
`outbox/`, and `.claude/skills/`). Your bundled skills live under `$AGENT_DIR/.claude/skills/`.
- `HIVE_ROOT` — the shared hive (`PROTOCOL.md`, the kanban, other agents).
At boot you also have `identity.md` (who you are) and `HIVE_ROOT/PROTOCOL.md` (the full coordination protocol). To message god or another agent, write ONE message JSON into `$AGENT_DIR/outbox/` (schema in PROTOCOL.md). When finished, send god an `"act":"done"` outbox message with a substantive result summary.
2. Temporal skills — concrete date ranges, relative to now
When your task is time-scoped, resolve the dates instead of computing them by hand. Each skill prints inclusive civil dates (`YYYY-MM-DD`, your local timezone) **and** the half-open `[startUtc, endExclusiveUtc)` instants for timestamp queries. All are read-only (clock + stdout only; no writes, no network).
Named shortcuts (invoke directly):
| Skill | Range it resolves | | --- | --- | | `/today`, `/yesterday` | the single civil day | | `/thisWeek`, `/lastWeek` | ISO week (Mon-start); this = Mon→today, last = prior full week | | `/last7Days`, `/last30Days` | rolling N-day window ending today | | `/thisMonth`, `/lastMonth` | this = 1st→today; last = prior full month | | `/thisQuarter`, `/lastQuarter` | this = quarter-start→today; last = prior full quarter | | `/thisYear`, `/lastYear` | this = Jan 1→today (YTD); last = prior full year |
For **any** window — including `last90Days`, `last12Months`, or arbitrary `lastNdays` / `lastNweeks` / `lastNmonths` — use `/temporal`, or call the resolver directly:
node "$AGENT_DIR/.claude/skills/temporal/when.mjs" last30Days # one window node "$AGENT_DIR/.claude/skills/temporal/when.mjs" # all windows node "$AGENT_DIR/.claude/skills/temporal/when.mjs" --json 90d # JSON only node "$AGENT_DIR/.claude/skills/temporal/when.mjs" --list # keywords
Convention: `this*` windows are **period-start → today** (to-date); `last*` named periods are the **full prior complete period**.
3. Integrations — via the loopback broker
External services are reached through the hive's **loopback broker**: a local, authenticated endpoint the harness runs on `127.0.0.1`. You call the broker; it holds the credentials and performs the outbound call on your behalf. You never see or store a secret, and only loopback callers are accepted — so integration access is brokered, auditable, and credential-free from your side.
How integrations are surfaced to you is **environment-dependent** — the exact set enabled depends on how you were spawned and the hive's configuration — so discover what's live at run time rather than assuming. The broker pattern is stable even as specific integrations are added; treat the list below as the current surface, not a fixed contract.
Concrete capabilities you may have right now (check availability before relying on one):
- **Reply to the originating Slack thread.** If your objective arrived from
Slack, your dispatch includes the exact loopback reply command — `node "<helper>" --channel <C> --thread <T> --text "<mrkdwn>"`. Use *that* command verbatim to post your result back to the thread. Send a substantive reply (a short `*bold*` headline + the actual outcome/links), never a bare "done".
- **Hive messaging.** Coordinate or hand off by writing an outbox message JSON
to god or another agent (`$AGENT_DIR/outbox/`). This is your always-available channel.
- **Semantic memory (MemPalace).** If enabled for you: `mempalace search
"<query>"` to recall shared knowledge, `mempalace wake-up` for a digest.
- **Enterprise Knowledge Graph.** If enabled: `node "$KG_CLI" search "<query>"`
for ranked passages, `node "$KG_CLI" list`, `node "$KG_CLI" get <id>` — use it for company-specific facts instead of guessing.
- **MCP integrations** (filesystem, git, and others) come pre-wired into your
session settings when enabled; invoke them as normal tools. The set is gated by the hive's consent configuration.
As additional brokered integrations land (calendar, mail, docs, web fetch, …), they follow the same shape: a brokered, credential-free call discoverable at run time. Pair them with the temporal skills above — resolve the date window first, then pass those concrete ISO bounds to the integration query.
4. Boundaries
- Temporal skills are **read-only** helpers — they never write or reach the
network. The broker mediates all external calls; you hold no credentials.
- Do **not** push or tag to any remote. Commit locally; **god** is the sole
integrator. Pause only for high-severity actions (remote push, paid/infra changes, deleting something you didn't create) — otherwise work autonomously.
- Finish by reporting to god (`"act":"done"`) with a real, substantive summary.
Local multi-agent harness for the terminal coding CLIs you already run — Claude Code, Antigravity (Gemini), OpenAI Codex, OpenCode, Crush, pi.dev, and GitHub Copilot CLI — with bring-your-own keys and local LLMs.
Repo: chaitanyagiri/munder-difflin
Other skills on munder-difflin.
- /last30Days
Resolve "last30Days" to a concrete ISO date range relative to your run time — a rolling 30-day window ending today. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a
Open skill - /last7Days
Resolve "last7Days" to a concrete ISO date range relative to your run time — a rolling 7-day window ending today. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a
Open skill - /lastMonth
Resolve "lastMonth" to a concrete ISO date range relative to your run time — the previous full calendar month. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a task
Open skill - /lastQuarter
Resolve "lastQuarter" to a concrete ISO date range relative to your run time — the previous full quarter. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a task over
Open skill - /lastWeek
Resolve "lastWeek" to a concrete ISO date range relative to your run time — the previous full week (Mon → Sun). Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a task
Open skill - /lastYear
Resolve "lastYear" to a concrete ISO date range relative to your run time — the previous full calendar year. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a task
Open skill

