agentleFS
Sign inStart free
iktok90-designiktok90-design/ai-smart-contract-auditor/skills/progress-and-streaming/SKILL.md

progress-and-streaming

Skill36 starsChanged 5 months agoMITView on GitHub

Convention for streaming progress on long-running audit commands. Use during /audit, /audit-deep, /audit-strict, /simulate, /exploit-chain — anything taking >10 seconds.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No spinners or animations (terminal output, not UI).

    line 44
  2. Very specific:

    No ANSI colors unless process.stdout.isTTY AND user hasn't disabled.

    line 45
  3. Distinctive:

    No hidden state — every meaningful event gets a line.

    line 46
  4. Distinctive:

    If the user interrupts (Ctrl+C / Esc), emit:

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Progress & streaming (meta-skill)

Long audits with no output feel broken. This skill defines the contract for live updates.

When to stream

  • Any command expected to take >10 seconds.
  • Any command that dispatches multiple subagents in parallel.
  • Any command that reads many files (>5).

What to emit

Stream short status lines to stdout as work progresses. Each line should be self-contained — a user reading the third line shouldn't need the first two.

Stages

[1/5] Inventory     scanning src/ for Solidity files
[1/5] Inventory     found 12 files (3,456 LoC), Solidity 0.8.24
[2/5] Skills        running reentrancy, access-control, oracle-manipulation, …
[2/5] Skills        18 skill checks complete (847ms)
[3/5] Specialists   dispatching: amm-specialist, attacker, defender (parallel)
[3/5] Specialists   amm-specialist:    3 findings
[3/5] Specialists   attacker:          5 findings (1 chain candidate)
[3/5] Specialists   defender:          2 missing-defense notes
[4/5] Consolidate   dedup + severity assignment
[5/5] Report        12 findings, grade C → wrote auditsentry-reports/audit-2026-05-13.md

Format

  • [N/M] STAGE message — N is current stage, M is total
  • Stage name padded to 12 chars for alignment
  • Message in lowercase, terse, no "I am now ..." prose
  • Numbers preferred over adjectives ("847ms" beats "quickly")

Don't

  • No spinners or animations (terminal output, not UI).
  • No ANSI colors unless process.stdout.isTTY AND user hasn't disabled.
  • No hidden state — every meaningful event gets a line.

Cancellation

If the user interrupts (Ctrl+C / Esc), emit:

[!]  cancelled at stage [3/5]: dispatched specialists are still running, partial results saved to .auditsentry/partial/

Implementation note

In a Claude Code slash-command context, this means: emit short text messages between tool calls. The user sees them streamed. Don't batch all updates into a final block.

Related

  • [[caching-and-incremental]] — show cache stats in stage [1/5]
  • [[confidence-scoring]] — show per-skill confidence as findings come in

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free