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/formin/spec-kit-harness/speckit.harness.exploregit clone --depth 1 https://github.com/formin/spec-kit-harnessWrote 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/commands/formin/spec-kit-harness/speckit.harness.explore)<a href="https://agentmods.dev/commands/formin/spec-kit-harness/speckit.harness.explore"><img src="https://agentmods.dev/badge/commands/formin/spec-kit-harness/speckit.harness.explore.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 5.1 | $0.00016 | $0.01334 |
| Opus 5 | $0.00008 | $0.00667 |
| Sonnet 5 | $0.00003 | $0.00267 |
| Haiku 4.5 | $0.00002 | $0.00133 |
Grade A, and why
speckit.harness.explore 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 6d 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 — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Budget-Aware Exploration Loop
Research the given question with a strict separation of concerns, following Harness-1 (arXiv:2606.02373):
- Policy (your reasoning) makes only semantic decisions: what to search, which candidate to inspect, what to retain at which importance, when to stop.
- Harness (the state files + the bookkeeping rules below) owns everything else: deduplication, compression, eviction, budget accounting.
Keep policy and bookkeeping separate. Spend your reasoning on search strategy and relevance judgments; execute bookkeeping mechanically, exactly as specified.
User Input
$ARGUMENTS
$ARGUMENTS is the research question for this session. If empty, use the
mission from budget.md. If both are empty, ask the user for a question and
stop.
Prerequisites
Resolve HARNESS_DIR as in /speckit.harness.init (feature directory →
harness/ subdirectory, else the configured fallback). If budget.md does
not exist there, run the initialization procedure from
/speckit.harness.init first (with $ARGUMENTS as the mission), then continue.
Steps
1. Render the working slice (never the full state)
Read configuration (config file → env vars, as in init) for budgets and slice sizes, then load ONLY:
budget.md: the budget table, stop conditions, and the last 5 action-log rows.curated.md: toprendering.curated_sliceentries, ordered critical → high → medium → low.candidates.md: up torendering.candidates_slicerows with statusneworinspected(the open frontier).observations.md: the lastrendering.observations_sliceentries.
This rendered slice — not your memory of earlier turns, and not the full
files — is your working state. If the slice plus your plan would exceed the
context_tokens cap, shrink the slices further, dropping low importance
entries first.
2. Iterate: decide → act → bookkeep
Repeat until a stop condition fires.
a. Policy decision (semantic). Based on the rendered slice, choose ONE:
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.
- 6d ago First seen · 127 lines · 16 tokens per session scan A c1ebf5291753
speckit.harness.explore is a command published in the GitHub repository formin/spec-kit-harness (5 stars, last pushed 2mo ago), licensed MIT. It adds 16 tokens to every session and 1,334 once invoked, about $0.0001 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 commands, from other repositories
sddp-autopilot
Command description: Run the full SDD feature-delivery pipeline. Argument hint: [optional: feature description; omit to select the first unchecked epic] Command category: orchestration Prerequisites: autopilot:enabled, product-document:planning-ready, technical-context:planning-ready.
sddp-implement-qc-loop
Command description: Run implement and QC in a continuous loop. Argument hint: [optional: feature directory or branch name] Command category: orchestration Prerequisites: spec, plan, tasks.
sddp-implement
Command description: Implement the current feature tasks. Argument hint: [optional: phase or task to start from] Command category: feature-delivery Prerequisites: spec, plan, tasks, checklists:complete-if-present.
sddp-amend
Command description: Propagate a bootstrap change across canonical project artifacts and the project plan. Argument hint: [project-level change to propagate across bootstrap artifacts] Command category: project-bootstrap Prerequisites: project-instructions, product-document, technical-context, project-plan.
sddp-checklist
Command description: Generate and evaluate a requirements quality checklist. Argument hint: [optional: quality focus or feature context] Command category: feature-delivery Prerequisites: spec, plan.
sddp-init
Command description: Initialize or amend project governance rules. Argument hint: [project description and principles] Command category: project-bootstrap Prerequisites: none.