Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/datit309/supergraph/handoffnpx skills add datit309/supergraph --skill handoffgit clone --depth 1 https://github.com/datit309/supergraphWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00044 | $0.00633 |
| Opus 5 | $0.00022 | $0.00316 |
| Sonnet 5 | $0.00009 | $0.00127 |
| Haiku 4.5 | $0.00004 | $0.00063 |
Grade A, and why
handoff scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/supergraph:handoff
Compress session state to a compact document. The next session picks up exactly where this one left off.
Announce: "📦 /supergraph:handoff — compacting session state..."
Steps
1. Capture current state
Gather live state — do not rely on memory:
git status --short
git log --oneline -10
Read active plan file (if any) from docs/supergraph/plans/.
2. Write handoff document
Save to $TMPDIR/supergraph-handoff-<YYYY-MM-DD>.md — NOT in the workspace (avoids accidental commits).
# Supergraph Handoff — <YYYY-MM-DD>
## Goal
[One sentence: what we were trying to accomplish]
## Status
[In-progress / Blocked / Ready for next phase]
## Active Plan
Path: [docs/supergraph/plans/<slug>.md]
Current task: Task N — [description]
Remaining tasks: [list with Status values]
## What Was Done
- [Completed task summaries — brief, no code]
## What Is Left
- [Remaining tasks, in order]
- [Any stuck tasks and why]
## Blockers / Open Questions
- [List — or "none"]
## Key Files Changed
- [path] — [what changed, one line]
## Git State
Branch: [branch name]
Last commit: [hash] [message]
Uncommitted: [list or "none"]
## Environment Context
[Copy from plan's Environment Context section]
## Suggested Skills for Next Session
1. /supergraph:scan (always first)
2. /supergraph:execute plan <slug> (continue tasks)
or /supergraph:tdd (if resuming single task)
3. /supergraph:fix → /supergraph:verify → /supergraph:review
## Notes
[Anything non-obvious a fresh agent needs to know]
Rules for content:
- Reference artifacts by path or URL only — never duplicate code or file content
- Keep the document under 150 lines
- No session-specific context (e.g., "as I said earlier") — self-contained
3. Report handoff location
✅ /supergraph:handoff complete
- Document: $TMPDIR/supergraph-handoff-<date>.md
- Status: [current status]
- Resume: /supergraph:scan → read handoff doc → /supergraph:execute plan <slug>
Rules
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 94 lines · 44 tokens per session scan A 072b45ff2403
handoff is a skill published in the GitHub repository datit309/supergraph (21 stars, last pushed 5d ago), licensed MIT. It adds 44 tokens to every session and 633 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
independent-review
Build the standardized prompt for a fresh reviewer subagent that evaluates implemented work against its spec without access to the implementation conversation. Use after an implementer subagent completes a spec slice (when the slice is ready for REVIEWED), or after a deviation log is written (for reconciliation…
pipeline
Use when the user wants a feature idea taken end-to-end in one autonomous run — phrases like "run the whole pipeline", "take this feature from idea to finished branch", "brainstorm then build it autonomously", "do everything from idea to merged", "implement all phases without stopping". Triggers when they want…
claude-delegate
Delegate an implementation or investigation task to Claude.
claude-rescue
Delegate an implementation or investigation task to Claude.
handoff-to-main-task
From a host-identified ephemeral Codex side conversation, preview and send a structured handoff to the uniquely verified main task, then open it. Fail closed outside that side-conversation boundary, for subagent coordination, or when the target is ambiguous.
create-handoff-prompt
Create a self-contained, copy-ready handoff prompt from the current conversation, preserving decisions, rationale, unresolved questions, next steps, work status, and authority boundaries. Do not send messages or manage Codex tasks.