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 commands/bitflight-devops/hallucination-detector/start-issuegit clone --depth 1 https://github.com/bitflight-devops/hallucination-detectorWhat 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 | $0.00000 | $0.00910 |
| Opus 5 | $0.00000 | $0.00455 |
| Sonnet 5 | $0.00000 | $0.00182 |
| Haiku 4.5 | $0.00000 | $0.00091 |
Grade A, and why
start-issue 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 — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Start Work on Issue
You are beginning work on GitHub issue $ARGUMENTS.
Follow these steps in order. Do not skip steps. Do not jump to implementation.
Step 1: Load the issue
node .claude/scripts/gh-api.cjs issue view $ARGUMENTS
Read the full issue body, labels, milestone, and assignee. Extract:
- Title and summary
- Impact type label (e.g.
impact: additive) - Risk level label (e.g.
risk: low) - Phase label (e.g.
phase: 1-additive-patterns) - Acceptance criteria from the issue body
If no acceptance criteria exist in the body, state that and ask the user whether to proceed or define criteria first.
Step 2: Load deep analysis and comments
node .claude/scripts/gh-api.cjs issue comment search $ARGUMENTS \
--section "Deep Analysis: Implementation Impact"
If found, extract:
- Dependencies (blocked by / blocks)
- Files touched
- Failure modes
- Contract impact
If no deep analysis comment exists, proceed using the issue body alone. If a blocking dependency is listed and still open, warn the user before proceeding.
Step 3: Determine workflow weight
Based on the labels extracted in Step 1:
- Lightweight (phase 1 + risk low + impact additive): Skip the research substep in Step 5. Go straight from understanding to objectives.
- Standard (everything else): Follow all substeps in Step 5.
Step 4: Create a feature branch
Derive a branch name from the issue number and title:
- Format:
feat/<issue-number>-<kebab-case-slug>(max 50 chars for the slug) - Example: issue #15 "Cognitive bias detection" →
feat/15-cognitive-bias-detection
Check if the branch exists first:
git branch --list "feat/$ARGUMENTS-*"
If it exists, check it out. Otherwise create it:
git checkout -b <branch-name>
Step 5: Plan the work
Follow the Working Process defined in .claude/CLAUDE.md § "Working Process" (steps 1–4: understand, research, objectives, gap analysis). If this is a lightweight issue (Step 3), skip the research substep.
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 · 132 lines · 0 tokens per session scan A ce7baae49a80
start-issue is a command published in the GitHub repository bitflight-devops/hallucination-detector (7 stars, last pushed 29d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 910 tokens. 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 commands, from other repositories
agt-status
Show the active AGT governance status for this Claude Code session.
thumbgate-blocked
Show what ThumbGate has actually blocked — gate enforcement stats and the full enforcement matrix. Use for "what has ThumbGate blocked", "show gate stats", "is enforcement working", "how many tokens did we save", "enforcement matrix".
capture-feedback
Quick feedback capture with structured signals.
check-gates
Run a Pre-Action Gate check against prevention rules before executing a risky action.
plan
Plan a task with scope contract and confidence check.
ng-context-pack
Portable command prompt generated from skills/briefing-an-agent/SKILL.md. Edit the skill, then run python tools/ng.py gen-commands; do not edit this file by hand.