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/davet47/hashloom/speckit.hashloom.gategit clone --depth 1 https://github.com/davet47/hashloomWrote 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/davet47/hashloom/speckit.hashloom.gate)<a href="https://agentmods.dev/commands/davet47/hashloom/speckit.hashloom.gate"><img src="https://agentmods.dev/badge/commands/davet47/hashloom/speckit.hashloom.gate.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 | $0.00024 | $0.01106 |
| Opus 5 | $0.00012 | $0.00553 |
| Sonnet 5 | $0.00005 | $0.00221 |
| Haiku 4.5 | $0.00002 | $0.00111 |
Grade A, and why
speckit.hashloom.gate 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 5d 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 — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hashloom Seam Gate
Runs the feature's seams through hashloom's cached verification. Invoked as
a hook by /speckit.implement (both before_implement and
after_implement), or directly by the user.
User Input
$ARGUMENTS
Determine the mode
- Invoked as a
before_implementhook → modeprecheck. - Invoked as an
after_implementhook → modegate. - Invoked directly: use
$ARGUMENTSif it saysprecheckorgate; default togate.
Run the script
-
Run
.specify/scripts/bash/check-prerequisites.sh --jsonfrom the repo root and parseFEATURE_DIR(absolute path). -
From the repo root, run:
.specify/extensions/hashloom/scripts/bash/hashloom-gate.sh <mode> <FEATURE_DIR> -
Parse the single JSON object it prints:
ok— the verdict (true/false)results.cached_pass— units proven green by the cache, no test ranresults.pass— units verified green just nowresults.fail—{name, summary}for failing unitsresults.error—{name, code, message}: unknown contract names, missing impl files (not yet implemented), impls without tests, ...spec_only— manifest names with no impl (types); nothing to runinferred— contracts still awaiting human review (advisory)note— present when there was nothing to gate (no/empty manifest)
Exit codes: 0 ok (including nothing-to-gate), 1 verification failed
(gate mode only), 2 infrastructure error (hashloom missing, index failed,
bad arguments — the JSON error.code says which).
Mode: precheck (before_implement)
The precheck NEVER blocks implementation; it harvests cache hits.
- If
noteis present (no seams), say so in one line and continue with implementation normally. - For every unit in
results.cached_pass: find the task(s) inFEATURE_DIR/tasks.mdwhose whole purpose is implementing or testing that unit (match on the unit name or its impl/test paths in the task description). Mark each such task[X]and append(cached-pass via hashloom — already verified). If a task covers more than the cached unit, leave it open and note the partial coverage. - Report a compact table: cached-pass (skipped work), pass, fail, error,
spec_only. Units in
fail/errorare simply work the coming tasks must do — do not treat them as blockers here. Typically, not-yet-implemented units appear as errors (missing impl file); that is expected before implementation. - If the script exited
2, report theerror.code/messageas a warning, tell the user the cache precheck was skipped, and continue. - Then proceed with implementation. While implementing, prefer the
contract as the spec for each seam, and after finishing a unit run
hashloom verify --radius <name>— green early beats red late.
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.
- 5d ago First seen · 95 lines · 24 tokens per session scan A b51a45bcfb04
speckit.hashloom.gate is a command published in the GitHub repository davet47/hashloom (2 stars, last pushed 3d ago), licensed Apache-2.0. It adds 24 tokens to every session and 1,106 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
spec-kitty.implement
Spec-Driven Development for serious software developers. Spec Coding with with Claude, Cursor, Gemini, Codex. Kanban dashboard, git worktrees, auto-merge and more.
bench-capture
Score the 2 bench variant folders for one size (build/acceptance/regression/conventions/capture + rubric + cross-solution review), record, report, and reset.
autospec.plan
Generate YAML implementation plan from feature specification.
OPSX: Continue
Continue working on a change - create the next artifact (Experimental).
VibeGuard: Live Truth
Verify live claims with fresh facts, inferences, and unresolved gaps.
kiro-validate-design
Command "kiro-validate-design" from gotalab/cc-sdd, covering technical design validation, core task, execution steps, important constraints and tool guidance.