sdlc-loop

A command that repeatedly runs eligible software-development tasks until the workflow reaches a gate, blocker, safety limit, or completion. SDLC means the software development life cycle.

In plain words
What is it for?
Use it to keep development tasks moving automatically, with options for workload tier, iteration limit, and a recurring schedule when the host supports scheduling.
Why use it?
It removes the need to manually prompt the agent to continue after each batch of work.

Command

Install

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.

agentmods
npx agentmods add commands/evolplus/talos/sdlc-loop
Clone the repo
git clone --depth 1 https://github.com/evolplus/talos
Per session 37 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,143 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00037 $0.01143
Opus 5 $0.00018 $0.00571
Sonnet 5 $0.00007 $0.00229
Haiku 4.5 $0.00004 $0.00114

Measured 2d ago against content hash 978ac64f0804, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

sdlc-loop scanned grade C 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 2d 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

4. Dispatch (Steps 4.5 availability, 4.6 journal+local detached worktree for code roles, 5 dispatch), then on each return: validate exit criteria and run the Step 7 finalization transaction: promote validated role-owned
commands/sdlc-loop.md · 36 lines

How it starts

The opening of the file, as written. The whole thing — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Orchestrator. Enter Autonomous Execution Loop mode per .claude/rules/autonomous-loop.md §15. Run continuously, dispatching eligible tasks yourself, until a halt condition fires — do NOT yield back to the operator after a single batch.

Arguments (may be empty): $ARGUMENTS

Setup (do this once, before the loop)

  1. Parse $ARGUMENTS:

    • --tier <aggressive|standard|conservative> → use this as the active workload tier for this run (overrides §13.3 env/SRS resolution for the session). If absent, resolve the tier normally per .claude/rules/workload-tier.md §13.3.
    • --max-iterations <N> → per-invocation soft cap (§15.6). Default 50.
    • --schedule [cadence] → after starting the loop, ALSO register a scheduled task using whatever scheduled-task mechanism the host provides (e.g. a create_scheduled_task tool in Cowork; cron / an SDK driver under bare Claude Code) whose prompt is exactly /sdlc-loop (carry forward --tier/--max-iterations but NOT --schedule, to avoid stacking schedules). Cadence default: hourly (0 * * * *). If no scheduler is available in this environment, say so and fall back to manual continuation rather than erroring. Confirm the registration to the operator with the task id + cadence. Note to the operator: each fire is a fresh run that resumes via §14; a fire that hits a human gate (H1–H4) will halt and re-fire fruitlessly until the gate is resolved, so they may pause the schedule for long-lived gates.
  2. Run the §9 pre-flight ONCE: Step 0 (git repo + committer identity — halt with NEEDS_CONTEXT if identity unset), Step 0.5 (root-level plan-update*.json straggler sweep), Step 0.6 (crash-recovery reconciliation of any surviving dispatch journal per .claude/rules/crash-recovery.md §14). Log the active tier in the start summary.

The loop (repeat until a halt condition)

Each iteration is one §9 cycle. Read project state from disk at the top of every iteration — never from a list you built earlier in this run (.claude/rules/autonomous-loop.md §15.2):

Read the full file on GitHub · 36 lines

Changes

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.

  1. 2d ago First seen · 36 lines · 37 tokens per session scan C 978ac64f0804

Subscribe to this mod's changes

sdlc-loop is a command published in the GitHub repository evolplus/talos (8 stars, last pushed 28d ago), licensed Apache-2.0. It adds 37 tokens to every session and 1,143 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.