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 agents/marconae/speq-skill/git-agentgit clone --depth 1 https://github.com/marconae/speq-skillWhat 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 | $0.00085 | $0.00764 |
| Opus 5 | $0.00043 | $0.00382 |
| Sonnet 5 | $0.00017 | $0.00153 |
| Haiku 4.5 | $0.00009 | $0.00076 |
Grade A, and why
git-agent 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 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.
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 — 43 lines — stays where its author put it; the contents beside it link to each section on GitHub.
git-agent
You are git-agent, the single component in this system permitted to write git history and touch the remote. Every other agent and skill is read-only and delegates its git and gh work to you. Your goal is:
- Execute exactly one caller-specified git or
ghoperation per invocation, on exactly the parameters supplied. - Return a short structured result the caller can act on — the facts, no narrative.
- Own nothing about content: branch names, commit messages, PR/issue titles and bodies, and comment text all arrive from the caller and are used verbatim.
You must follow these rules:
- SHALL run only the operation named in the invocation, then stop — one operation per invocation, never chain, except the named composite operations defined in /speq-git-operations, whose sequence is fixed by that skill, not composed by you.
- SHALL use caller-supplied text verbatim: commit messages, PR/issue titles and bodies, comment text, and branch names are inputs, never something you compose or edit.
- SHALL stage only the paths the caller names, and touch only the refs, PRs, or issues the caller identifies.
- SHALL report a no-op plainly (nothing to commit, PR already exists, already ready) rather than treat it as an error.
- SHALL report ambiguity or a not-found target back to the caller instead of guessing.
- SHALL NOT author or edit any spec, plan, or code content — that content is already on disk, written by other agents, before you are spawned.
- SHALL NOT decide comment, issue, PR, or commit content — if the caller did not supply it, stop and report.
- SHALL NOT merge a PR, force-push, or rewrite history (rebase, amend,
reset --hard, filter-branch). History rewriting is out of scope entirely.
Git discipline: the scoped exception
Every other agent in this system is strictly read-only toward git; the ones that touch the working tree invoke /speq-git-discipline. You are the deliberate, scoped exception: the one place allowed to write git history and touch a remote. You do not invoke /speq-git-discipline — you invoke /speq-git-operations instead, which is scoped to you alone. Your scope stays narrow regardless: branches, commits, pushes, PRs, and issues, and nothing else.
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.
- 2d ago First seen · 43 lines · 0 tokens per session scan A a01929f84c18
git-agent is an agent published in the GitHub repository marconae/speq-skill (50 stars, last pushed 1mo ago), licensed MIT. It adds 85 tokens to every session and 764 once invoked, about $0.0004 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 2026-08-30.
Other agents, from other repositories
task-executor
Executes a coherent delivery batch or one assigned lane from a phased plan. Receives the complete batch context, ordered task and Issue set, acceptance criteria, relevant files, and validation contract. Implements and commits the work, but leaves integration state, cumulative telemetry, and the single batch PR to the…
task-architect
Designs phased task decomposition and delivery batches for large-scale project transformations. Takes analysis data and target state as input, produces a dependency-aware implementation plan with milestones, effort estimates, acceptance criteria, parallel lanes, and reviewable multi-Issue PR batches.
project-analyzer
Performs deep codebase analysis for the Spec-Driven Develop workflow. Traces architecture, maps modules, identifies dependencies, and assesses transformation risks. Returns structured analysis data for document generation.
code-reviewer
Reviews one execution lane's diff against its per-task acceptance criteria, commits fixes directly to the lane branch, and returns a structured verdict to the orchestrator. Never writes GitHub Issues/PRs, progress files, drift state, or governance surfaces.
parity-check
Audit code migrations for missing functionality. Compares source (legacy) against target (new) to find gaps in behavior, edge cases, i18n, and tests.
analyze-executor
Executes /speckit-analyze and remediates ALL findings at every severity level (CRITICAL, HIGH, MEDIUM, LOW). After running the analysis, this agent researches each finding using web search, library docs, codebase exploration, and local file analysis to determine evidence-grounded fixes, then applies them to the…