Content
73%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body delivers a well-sequenced, safety-conscious workflow with strong validation checkpoints and highly concrete commands for both host and Bitterbot tooling. Its main weakness is redundancy — several rules are restated two or three times — and a couple of command syntaxes are left elided.
Suggestions
State each core rule once: remove the duplicate model self-check (Step 0 repeats a Core rule verbatim) and consolidate the triple-repeated 'Bitterbot does not change host firewall/SSH/OS updates' warning into a single place.
Give the full 'bitterbot cron add' syntax (all flags for cadence, time window, and output location) instead of the elided 'bitterbot cron add --name <name> ...', since scheduling is a core advertised capability.
Fix or remove the 'docs/reference/templates/TOOLS.md' reference — it points outside the skill bundle — and consider moving the memory-write conventions to a reference file to shorten the main body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and free of concept explanations, but contains clear repetition: the model self-check appears in both 'Core rules' and 'Step 0', and the warning that Bitterbot never changes host firewall/SSH/OS updates is stated three times (Core rules, Step 2, and 'Bitterbot command accuracy'). It fits 'mostly efficient but could be tightened' better than the 'minor instances of over-explanation' level above. | 3 / 5 |
Actionability | Concrete, copy-paste-ready commands appear throughout with per-OS variants ('ss -ltnup', 'ufw status', 'bitterbot security audit --deep', exact cron job names), but 'bitterbot cron add --name <name> ...' elides its actual flags and no example hardening (firewall/SSH) commands are provided. | 4 / 5 |
Workflow Clarity | Steps 0-8 are clearly sequenced with read-only checks first, the plan shown before any change, per-step confirmation with 'Stop on unexpected output and ask for guidance', an explicit verification pass (Step 8 re-checks firewall, ports, remote access, and re-runs the audit), and an always-required confirmations checklist. | 5 / 5 |
Progressive Disclosure | No bundle files exist, and the single document is well-sectioned with a flat, navigable structure. Minor gaps: the reference to 'docs/reference/templates/TOOLS.md' points outside the skill bundle, and the memory-write conventions are a candidate for a separate reference file. | 4 / 5 |
Total | 16 / 20 Passed |