BC claude-usage-companion
Monitor your Claude programmatic credit pool and align your interactive window — on your always-on box. Since June 2026 Anthropic meters non-interactive usage (Claude Agent SDK, `claude -p`, Claude Code GitHub Actions, third-party agents) from a SEPARATE monthly credit pool ($20 Pro / $100 Max 5x / $200 Max 20x, API rates, no rollover) instead of your 5-hour interactive window. claude-usage-companion (1) tracks that credit pool — month-to-date burn, projected month-end spend, and an alert before you run dry so automations don't stop unexpectedly; and (2) reminds you, the human, at your anchor time to send your first interactive prompt so your 5-hour window lines up with your working day. Use this skill when the user mentions Claude usage limits, the programmatic/Agent-SDK credit pool, running out of credits, monitoring claude -p or GitHub Actions spend, ccusage, usage reports, budgeting Claude automation on a VPS, or wants to align/start their interactive session window. It only reports and reminds — it never automates the model or circumvents any limit.
Monitor your Claude programmatic credit pool and align your interactive window — on your always-on box.
As a process C 59/100 · Has gaps — weak spots: result and completion, failures and branches, running it twice
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
How to improve
- Shorten the description to 1024 characters.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 6
✓ No critical or high findings
Medium and low: 6
-
medium Dangerous commands
cmd-persistencebin/usage-companion.mjs:244Persistence mechanism (cron / launchd / scheduled task / autorun registry) (string literal in code, not executed)console.log(`# ( crontab -l 2>/dev/null; node ${self} install --cron-only ) | crontab -`);code literal -
low Dangerous commands
cmd-cron-mentionbin/usage-companion.mjs:239Mentions editing / listing crontab (string literal in code, not executed)console.log("# claude-usage-companion — paste into `crontab -e`");code literal -
low Dangerous commands
cmd-cron-mentionbin/usage-companion.mjs:244Mentions editing / listing crontab (string literal in code, not executed)console.log(`# ( crontab -l 2>/dev/null; node ${self} install --cron-only ) | crontab -`);code literal -
low Dangerous commands
cmd-cron-mentiondocs/index.html:153Mentions editing / listing crontabcrontab -e <span class="cm"># paste it, save</span></pre></div>
-
low Dangerous commands
cmd-cron-mentionREADME.md:73Mentions editing / listing crontabcrontab -e # paste it, save
-
low Dangerous commands
cmd-cron-mentionSKILL.md:68Mentions editing / listing crontabcrontab -e # paste it, save
Files scanned: 6. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1070 chars, limit 1024
Process rating: all ten parameters 59/100
- 0Result and completion. Does not say what the result is
- 0Failures and branches. Linear process with no failure handling
- 30Running it twice. 2 mutating operations with no state check
- 60Tools and files. Uses tools (bash, node) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 11 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 1035 tokens
- 100Progress reporting. Reports progress
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1070: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +4Structure: 7 headings
- +3Step-by-step instructions: 11 items
- +4Has examples (3 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 56.