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.
git clone --depth 1 https://github.com/alfredoperez/speckit-companionWrote 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/alfredoperez/speckit-companion/speckit.companion.classify)<a href="https://agentmods.dev/commands/alfredoperez/speckit-companion/speckit.companion.classify"><img src="https://agentmods.dev/badge/commands/alfredoperez/speckit-companion/speckit.companion.classify/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/alfredoperez/speckit-companion/speckit.companion.classify"><img src="https://agentmods.dev/badge/commands/alfredoperez/speckit-companion/speckit.companion.classify.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00021 | $0.00520 |
| Opus 5.5 | $0.00008 | $0.00208 |
| Sonnet 5.5 | $0.00004 | $0.00104 |
| Haiku 4.5 | $0.00002 | $0.00052 |
Grade A, and why
speckit.companion.classify 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 20d 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.
What it actually says
Classify Change Size
Emit a single complexity signal — simple, normal, or oversized — that the Companion
workflow's routing step reads to right-size the pipeline. On the workflow path there is no
complexityFastPath on/off setting: the thresholds live here, in the workflow, not in a
VS Code toggle.
This is a thin, read-only step. It does not write .spec-context.json and does not create or
edit any spec files; it only reports a size.
Heuristic (thresholds live here, not in a setting)
Estimate the scope of the change from spec.md (and plan.md/tasks.md if they already exist):
- small: the change plausibly touches ≤ 5 files and decomposes into ≤ 10 tasks, and is easy to undo and easy to check. A change that is hard to reverse (a migration, deleted stored data, a published contract) or hard to verify is never small, whatever its file count.
- oversized: the change clearly exceeds the small bar by a wide margin (broad multi-subsystem work, many new files, or a long task list).
- normal: anything in between (the default).
When unsure, prefer normal — the routing step's safe default is the full pipeline, so an
ambiguous estimate never skips a phase.
Output
Print exactly one line so the size is visible in the run log:
[companion] size=<simple|normal|oversized>
Expose the same value as structured output size (so a switch node can read
steps.classify.output.size). Routing contract:
simple→ the workflow folds toward implement (less ceremony).simpleis the verdict every reader of the recorded size expects;smallnames the bar, never the verdict.oversized→ the workflow prints a visible warning and still runs the full pipeline. It never silently skips a phase.normal(and any unresolved value) → the full pipeline.
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.
- 20d ago Changed 019bb6076e90
- 1mo ago Changed · -2 lines db0fa480cf83
- 1mo ago First seen · 48 lines · 21 tokens per session scan A 37fc3fef610b
speckit.companion.classify is a command published in the GitHub repository alfredoperez/speckit-companion (99 stars, last pushed 10d ago), licensed MIT. It adds 21 tokens to every session and 520 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-09-03.
Other commands, from other repositories
peachflow:implement
Execute implementation of the current sprint's tasks. Works within a sprint worktree, plans approach, implements tasks in order, and tracks progress in the graph. Uses plan mode for each task.
peachflow:create-sprint
Find unlocked tasks in the current quarter, create a new sprint with up to 10 tasks, and set up a git worktree for implementation. Starts the sprint workflow.
peachflow:discover
Start product discovery for new projects OR add new features/initiatives to existing projects. Creates BRD/PRD and populates epics in the graph.
peachflow:plan
Break down epics into user stories and tasks using Claude's native plan mode. Explores context, writes plan for approval, then creates work items.
peachflow:graph
Launch a local visualization server to view the project's work item graph. Shows epics, user stories, tasks, and their relationships in an interactive browser-based interface.
peachflow:clarify
Review and resolve pending clarifications stored in the graph. Clarifications are questions that arose during discovery or planning.