Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

/launch - Session Start · majiayu000 | DecimalAI

majiayu000/launch

NewSkillScoreMaintained

Start a Space-Agents session. Displays welcome screen with project status.

agent-toolingby majiayu000

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Project Management pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/majiayu000-launch/SKILL.md

decimalai skills pull majiayu000-launch --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/majiayu000-launch/SKILL.md https://app.decimal.ai/s/majiayu000-launch/SKILL.md

Raw file: /s/majiayu000-launch/SKILL.md ↗ · machine-readable summary at /s/majiayu000-launch.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 19d ago

—-3% tokens0% turns23% of whole cases pass with skill (5/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	30%	0%
case-15	✗→✓	▲ Improved	-32%	0%
case-16	✗→✓	▲ Improved	-15%	0%
case-18	✗→✓	▲ Improved	-19%	0%
case-10	✓→✗	▼ Worse	183%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

majiayu000 / claude-skill-registry

skills/agent/launch/SKILL.md

View on GitHub ↗

License MITImported 8/10/2026Version v1

Skill contents

SKILL.md

/launch - Session Start

You are HOUSTON, the Flight Director. Calm, professional, NASA-style.

Process

Run bd list --tree to check state
If output contains "no beads database found" → display welcome screen with "not initialized" placeholders
Otherwise, query status:

bash bd ready # Unblocked work bd stats # Counts

Read last CAPCOM entry: grep -n "^## \[" .space-agents/comms/capcom.md | tail -1 then read from that line
Display welcome screen

Welcome Screen

Only output the welcome screen. All context goes in {briefing}.

┌────────────────────────────────────────────────────────────────┐
│  ███████╗██████╗  █████╗  ██████╗███████╗                      │
│  ██╔════╝██╔══██╗██╔══██╗██╔════╝██╔════╝                      │
│  ███████╗██████╔╝███████║██║     █████╗                        │
│  ╚════██║██╔═══╝ ██╔══██║██║     ██╔══╝                        │
│  ███████║██║     ██║  ██║╚██████╗███████╗                      │
│  ╚══════╝╚═╝     ╚═╝  ╚═╝ ╚═════╝╚══════╝                      │
│           █████╗  ██████╗ ███████╗███╗   ██╗████████╗███████╗  │
│          ██╔══██╗██╔════╝ ██╔════╝████╗  ██║╚══██╔══╝██╔════╝  │
│          ███████║██║  ███╗█████╗  ██╔██╗ ██║   ██║   ███████╗  │
│          ██╔══██║██║   ██║██╔══╝  ██║╚██╗██║   ██║   ╚════██║  │
│          ██║  ██║╚██████╔╝███████╗██║ ╚████║   ██║   ███████║  │
│          ╚═╝  ╚═╝ ╚═════╝ ╚══════╝╚═╝  ╚═══╝   ╚═╝   ╚══════╝  │
├────────────────────────────────────────────────────────────────┤
│            HOUSTON online. All systems nominal.                │
├────────────────────────────────────────────────────────────────┤
│  Project: {project}                                            │
│  Features: {feature_count} | Tasks: {task_count} | Bugs: {bugs}│
├────────────────────────────────────────────────────────────────┤
│  COMMANDS                                                      │
│    /launch          Start session (you are here)               │
│    /exploration     Analyze and plan                           │
│      brainstorm       Explore ideas → brainstorm reports       │
│      plan             Structure work → plan.md & Beads         │
│      review           Code review → bugs/tasks                 │
│      debug            Investigate issues → bugs                │
│    /mission         Execute from Beads                         │
│      solo             Direct execution (small, 1-3 tasks)      │
│      orchestrated     Agents-per-task (medium, 4-10 tasks)     │
│      ralph            Automatic background (large, 10+ tasks)  │
│    /capcom          Check status and progress                  │
│    /land            End session, save to CAPCOM                │
├────────────────────────────────────────────────────────────────┤
│  TREE                                                          │
│  {tree}                                                        │
├────────────────────────────────────────────────────────────────┤
│  READY                                                         │
│  {ready}                                                       │
├────────────────────────────────────────────────────────────────┤
│  BRIEFING                                                      │
│  {briefing}                                                    │
└────────────────────────────────────────────────────────────────┘

Placeholders

{project}: Epic title from bd list --tree
{feature_count}, {task_count}, {bugs}: From bd stats
{tree}: Output of bd list --tree (indent with │ )
{ready}: Output of bd ready or "No unblocked tasks"
{briefing}: Summary from last CAPCOM entry. If nothing: "All quiet. Ready for orders."

Illustrative Want to see what majiayu000/launch would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/2/2026, 05:51 UTC · 22 cases

Measured on 8 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

0%

23% vs 23%

Agent turns

0%

1.0 vs 1.0

Tokens

-3%

1,541.5 vs 1,586.25

fail → pass · 4pass → pass · 1fail → fail · 13pass → fail · 4

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I'm ready to kick off our current development session. Please run the flight director startup routine to check project state, gather active task statistics, pull the latest communication log briefing, and present the main session start welcome screen.	fail→fail	12,652	9,401	-26%	1	1	0%	1,304	1,215	-7%	0	0	—
case-02 /launch - Please start my flight director dashboard session, inspect the task tree and system readiness, fetch recent Capcom notes, and output the standard startup banner.	fail→fail	11,484	9,520	-17%	1	1	0%	1,843	1,177	-36%	0	0	—
case-03 Can you initiate HOUSTON's mission start protocol? Check the project status, count active work items, review the latest Capcom log summary, and print the session control ASCII box interface.	fail→fail	21,268	4,125	-81%	1	1	0%	2,296	1,304	-43%	0	0	—
case-04 I want to initialize our work environment. Should I run git status or npm test first to verify the repository layout?	fail→fail	11,738	11,117	-5%	1	1	0%	1,119	1,416	+27%	0	0	—
case-05 When running the workspace inspector, `bd list --tree` outputs 'no beads database found'. Should I run `bd init` to build a standard database file immediately?	fail→fail	6,158	4,573	-26%	1	1	0%	1,065	1,199	+13%	0	0	—
case-06 During session launch, should I read the Capcom communications log before verifying the task database state?	fail→pass	10,096	10,944	+8%	1	1	0%	1,492	1,939	+30%	0	0	—
case-07 To find work that is ready to be claimed without impediments, should I query `git status` or inspect issues on GitHub?	fail→fail	8,938	17,525	+96%	1	1	0%	1,499	1,433	-4%	0	0	—
case-08 We need numerical metrics on pending features, bugs, and tasks. Should we scan the filesystem markdown files or run a CLI counter?	fail→fail	13,978	6,039	-57%	1	1	0%	2,008	1,466	-27%	0	0	—
case-09 We need to parse the last entry header from our communications log. Should we read the entire file into memory using standard node fs or find the line number via command line tools?	fail→fail	17,687	17,612	-0%	1	1	0%	2,320	1,419	-39%	0	0	—
case-10 When extracting the final entry header line number from `.space-agents/comms/capcom.md`, should we pipe grep output into `head -1` or `tail -1`?	pass→fail	4,016	13,716	+242%	1	1	0%	565	1,599	+183%	0	0	—
case-11 When displaying the task database hierarchy during startup, should I pass `--recursive` or `--tree` to `bd list`?	fail→fail	13,926	12,371	-11%	1	1	0%	1,269	1,524	+20%	0	0	—
case-12 To isolate only the single latest header line number from the grep results on the Capcom log, should we pipe to `sort -u` or `tail -1`?	pass→pass	9,844	10,553	+7%	1	1	0%	867	1,715	+98%	0	0	—
case-13 To populate feature, task, and bug counts for the session dashboard, should we count lines in `.space-agents/` or run `bd stats`?	fail→fail	12,746	1,917	-85%	1	1	0%	1,257	1,308	+4%	0	0	—
case-14 When launching the session interface, should I print conversational text summaries before and after rendering the dashboard box?	fail→fail	12,890	2,435	-81%	1	1	0%	2,027	1,324	-35%	0	0	—
case-15 Where does the project name value come from when filling the `{project}` placeholder in the launch banner?	fail→pass	16,660	2,296	-86%	1	1	0%	1,971	1,335	-32%	0	0	—
case-16 In what sequence should database status checks and communication log reads occur during session initialization?	fail→pass	13,373	18,378	+37%	1	1	0%	1,892	1,617	-15%	0	0	—
case-17 Where are communication log records stored for the project agent communications channel?	fail→fail	10,011	10,816	+8%	1	1	0%	1,701	1,280	-25%	0	0	—
case-18 How are `{feature_count}`, `{task_count}`, and `{bugs}` populated in the launch interface?	fail→pass	11,060	12,269	+11%	1	1	0%	2,045	1,647	-19%	0	0	—
case-19 How do I create a new bug report or task item in the beads issue tracker CLI?	pass→fail	8,918	15,527	+74%	1	1	0%	1,692	1,356	-20%	0	0	—
case-20 What command marks a completed work item as closed in the beads task database?	pass→fail	11,261	12,872	+14%	1	1	0%	1,000	1,538	+54%	0	0	—
case-21 How do I view the detailed description and dependencies of a single specific work item in beads?	pass→fail	12,247	7,229	-41%	1	1	0%	1,241	1,452	+17%	0	0	—
case-22 When searching for tasks that are currently unblocked and ready for work, which CLI command should be executed?	fail→fail	7,403	8,031	+8%	1	1	0%	1,139	1,447	+27%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 8 counted toward the lift figure. The other 14 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of 0 percentage points is the difference between those two pass rates over the 8 comparable cases. 4 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 22.7%
Pass rate with the skill: 22.7%
Cases attempted: 22
Cases counted toward lift: 8
Cases that regressed: 4
Skill version measured: v1
Run completed: 9/2/2026, 05:51 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-06

Prompt

During session launch, should I read the Capcom communications log before verifying the task database state?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+98% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 19 days ago

Other measured skills in the registry, with their headline benchmark lift.