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/wasulajr/headsup/sflnpx skills add wasulajr/headsup --skill sflgit clone --depth 1 https://github.com/wasulajr/headsupWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/wasulajr/headsup/sfl)<a href="https://agentmods.dev/skills/wasulajr/headsup/sfl"><img src="https://agentmods.dev/badge/skills/wasulajr/headsup/sfl.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00121 | $0.01404 |
| Opus 5 | $0.00060 | $0.00702 |
| Sonnet 5 | $0.00024 | $0.00281 |
| Haiku 4.5 | $0.00012 | $0.00140 |
Grade A, and why
sfl 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 3d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
sfl: Save For Later (and resume)
Two modes. Default is save. If the invocation says resume (e.g. /sfl resume, or the launch prompt /nil injects), run resume instead.
SAVE mode
Checkpoint the current window so any fresh session can pick it up with zero re-derivation. One required write, one optional write, then a banner.
1. Durable checkpoint (optional)
If you maintain a memory system (project memory files, a notes directory, anything persistent across sessions), write the fuller checkpoint there: what was done, what is pending, exact open questions, file and resource paths, decisions made. Convert relative dates to absolute. Distill; do not dump the transcript. Reference that file in the entry's gov_memory field in step 2.
If there is no memory system, skip this step. The entry file from step 2 is then the complete checkpoint, and gov_memory: is (none).
2. Live per-window entry (what /nil consumes)
Resolve this window's identity:
~/.claude/sfl/lib/window-id.sh
It prints LABEL=, SLUG=, CWD=, STAMP=. The entry file is ~/.claude/sfl/<SLUG>.md: one file per window, and you overwrite it (the newest sfl per window wins; that is the whole point, so do not append or version it).
Write it through the Bash helper, NOT the Write tool. Claude Code guards direct Write-tool access to its own config directory (~/.claude/), which prompts on every save. The helper sfl-entry.sh writes via the Bash tool instead (covered by the allow rules setup.sh installs), so checkpoints save silently. Pipe the composed markdown to it via a heredoc, substituting <SLUG> and filling every field:
cat <<'SFLEOF' | ~/.claude/sfl/lib/sfl-entry.sh write '<SLUG>'
---
window: <LABEL>
project: <human project name>
cwd: <CWD>
saved_at: <STAMP>
gov_memory: <path to the durable memory file from step 1, or "(none)">
---
## Checkpoint
<2-4 sentences: what was being worked on, decisions made, what is pending, open questions. Distilled; the fuller version, if any, lives in gov_memory.>
## How to restart
<the single concrete first action on resume, plus any 2nd/3rd step. Concrete enough to act on immediately: "run X against Y", not "continue the work".>
SFLEOF
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.
- 3d ago First seen · 88 lines · 121 tokens per session scan A 53a28bff8f8e
sfl is a skill published in the GitHub repository wasulajr/headsup (1 stars, last pushed 16d ago), licensed MIT. It adds 121 tokens to every session and 1,404 once invoked, about $0.0006 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 2026-08-31.
Other skills, from other repositories
agent-context-isolation
Prevent agent output from polluting the main context window.
context-engineering
Guidance for designing, auditing, and improving context supplied to LLM agents, prompts, tools, RAG/MCP servers, and multi-agent systems. Use when the task involves context selection, prompt or agent architecture, tool descriptions, retrieval context, memory, compaction, autonomy, or evaluation of LLM behavior.
mot
System health check (MOT) for skills, agents, hooks, and memory.
no-task-output
TaskOutput floods the main context window with agent transcripts (70k+ tokens).
recall-reasoning
Search past reasoning for relevant decisions and approaches.
recall
Query the memory system for relevant learnings from past sessions.