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/rhuss/cc-spex/speckit.spex-collab.phase-splitgit clone --depth 1 https://github.com/rhuss/cc-spexWrote 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/rhuss/cc-spex/speckit.spex-collab.phase-split)<a href="https://agentmods.dev/commands/rhuss/cc-spex/speckit.spex-collab.phase-split"><img src="https://agentmods.dev/badge/commands/rhuss/cc-spex/speckit.spex-collab.phase-split.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 | $0.00007 | $0.03502 |
| Opus 5 | $0.00003 | $0.01751 |
| Sonnet 5 | $0.00001 | $0.00700 |
| Haiku 4.5 | $0.00001 | $0.00350 |
Grade B, and why
speckit.spex-collab.phase-split scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks the agent to reveal its instructionsmediumSystem prompt leakage
Directions to print, repeat or translate the system prompt extract configuration the operator did not intend to expose.
When single-phase mode is active (either threshold-defaulted or user-selected "Single phase (no split)"), output instructions that run implementation straight through all tasks with phase-manager called only once at the How it starts
The opening of the file, as written. The whole thing — 369 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Phase Split Proposal
Ship Pipeline Guard
If .specify/.spex-state exists and its mode is "ship" or status is "running", skip the phase split entirely and return immediately without prompting.
if [ -f ".specify/.spex-state" ]; then
MODE=$(jq -r '.mode // empty' .specify/.spex-state 2>/dev/null)
STATUS=$(jq -r '.status // empty' .specify/.spex-state 2>/dev/null)
if [ "$MODE" = "ship" ] || [ "$STATUS" = "running" ]; then
echo "Ship mode active, skipping phase split"
fi
fi
If ship mode is detected, return immediately. Do not display any proposal or prompt the user.
Extension Enabled Check
Verify spex-collab is active:
if [ ! -f ".specify/extensions/spex-collab/extension.yml" ]; then
echo "spex-collab extension not found, skipping"
fi
If the extension is not found, return without action.
Check for Existing Phase Plan
If .specify/.spex-state already has a collab.phase_plan with entries, a phase plan was previously confirmed:
EXISTING_PLAN=$(jq -r '.collab.phase_plan // empty' .specify/.spex-state 2>/dev/null)
If a phase plan exists and has entries:
- Display: "A phase plan already exists from a previous session."
- Show the existing plan as a table
- Ask: "Use existing plan, or create a new one?"
- If user chooses existing: return (implementation proceeds with saved plan)
- If user chooses new: continue with detection below
Resolve Spec Directory and Read Tasks
PREREQ=$(.specify/scripts/bash/check-prerequisites.sh --json --require-tasks --include-tasks 2>/dev/null)
FEATURE_DIR=$(echo "$PREREQ" | jq -r '.FEATURE_DIR')
Read tasks.md from FEATURE_DIR.
Estimate File Count
Estimate how many files the feature will touch. This determines whether to propose multi-phase splitting or silently default to single phase.
Parse File Paths from plan.md
If plan.md exists in FEATURE_DIR, extract file path references:
PLAN_FILE="${FEATURE_DIR}/plan.md"
ESTIMATED_FILES=0
ESTIMATION_SOURCE="heuristic"
if [ -f "$PLAN_FILE" ]; then
# Extract file paths: match patterns like src/foo/bar.sh, `path/to/file.md`, ./relative/path.yml
# Filter: must contain a directory separator (/), have a file extension (1-10 chars)
# Exclude: URLs (http/https), version-like patterns (v1.5)
FILE_PATHS=$(grep -oE '[a-zA-Z0-9_./-]+\.[a-zA-Z]{1,10}' "$PLAN_FILE" 2>/dev/null | \
grep '/' | \
grep -vE '^https?://' | \
grep -vE '^//' | \
grep -v '^v[0-9]' | \
sort -u)
FILE_COUNT=$(echo "$FILE_PATHS" | grep -c '.' 2>/dev/null || echo "0")
if [ "$FILE_COUNT" -ge 5 ]; then
ESTIMATED_FILES=$FILE_COUNT
ESTIMATION_SOURCE="plan.md"
fi
fi
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 · 369 lines · 7 tokens per session scan B 8551368749cf
speckit.spex-collab.phase-split is a command published in the GitHub repository rhuss/cc-spex (114 stars, last pushed 17d ago), licensed Apache-2.0. It adds 7 tokens to every session and 3,502 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it B with 1 finding (asks the agent to reveal its instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.