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 skills/thegreencedar/codestory/codestory-groundingnpx skills add TheGreenCedar/CodeStory --skill codestory-groundinggit clone --depth 1 https://github.com/TheGreenCedar/CodeStoryWrote 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/skills/thegreencedar/codestory/codestory-grounding)<a href="https://agentmods.dev/skills/thegreencedar/codestory/codestory-grounding"><img src="https://agentmods.dev/badge/skills/thegreencedar/codestory/codestory-grounding.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 | $0.00044 | $0.01427 |
| Opus 5 | $0.00022 | $0.00714 |
| Sonnet 5 | $0.00009 | $0.00285 |
| Haiku 4.5 | $0.00004 | $0.00143 |
Grade A, and why
codestory-grounding 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 3d 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 — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CodeStory Grounding
CodeStory keeps a local repository map and broad-search index so agents can reach useful evidence without rediscovering the same code every turn.
The target is always the repository being grounded. Pass its exact absolute
root as project on every CodeStory call. Never rely on a global active
workspace.
Direct Tool Loop
Call the tool that matches the task. Do not call status first.
Using this skill does not require an MCP call when the requested work is fully
local to an already named evidence surface—for example, inspecting or editing
the content of assets/desk.svg. Inspect that surface directly. Naming a path
does not make the evidence surface complete when the task asks about ownership,
dependencies, runtime behavior, architecture, change impact, or another claim
whose evidence extends beyond the file. For those tasks, select the narrowest
CodeStory tool that can add evidence; do not call broad ground as a pre-edit
ceremony.
- Resolve the target repository root.
- Call the intended tool with
project=<absolute-root>. - If the result says
state=preparingorstate=updatingand includesretry_after_ms, wait for that delay and retry the same tool with the same arguments. The delay tracks observed preparation progress, so honor the reported value instead of a fixed poll interval. Do not poll status or ask the user to set up CodeStory. - Preserve cited anchors in source claims. Read focused source only for the remaining evidence gaps.
CodeStory prepares its local repository map and shared per-user retrieval server
automatically. status and
the project-bound codestory://status{?project} resource are optional
diagnostics for a failed or unexpectedly slow request, not prerequisites for
normal grounding.
If CodeStory tools are hidden and deferred discovery is available,
search only for the intended tool, for example codestory mcp packet, then call
it directly. If the plugin MCP is unavailable, use ordinary source inspection
and report the visibility gap. Do not substitute CLI diagnostics for a live
plugin result unless the user explicitly asks.
What ships with it
25 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 675 B
- references/affected.md 3.7 KB
- references/bookmark.md 1.5 KB
- references/cache.md 1.9 KB
- references/context.md 1.5 KB
- references/doctor.md 3.1 KB
- references/drill-suite.md 5.7 KB
- references/drill.md 6.7 KB
- references/explore.md 3.0 KB
- references/files.md 1.9 KB
- references/generated-cli-syntax.md 2.3 KB
- references/generated-mcp-syntax.md 3.2 KB
- references/ground.md 2.2 KB
- references/index.md 4.2 KB
- references/packet.md 4.8 KB
- references/query.md 2.1 KB
- references/retrieval-rollout.md 3.8 KB
- references/search.md 3.7 KB
- references/serve.md 3.5 KB
- references/snippet.md 2.4 KB
- references/status-contract.md 7.6 KB
- references/symbol.md 1.1 KB
- references/trail.md 2.0 KB
- scripts/setup.ps1 6.5 KB runs code
- scripts/setup.sh 4.1 KB runs code
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.
- 3d ago First seen · 121 lines · 44 tokens per session scan A 3f2cf1908125
codestory-grounding is a skill published in the GitHub repository TheGreenCedar/CodeStory (37 stars, last pushed 4d ago), licensed Apache-2.0. It adds 44 tokens to every session and 1,427 once invoked, about $0.0002 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 skills, from other repositories
adding-a-language
How a new language enters Kivgraph's graph - the two routes and why one of them is almost always wrong, the four decisions that cannot be migrated later, the branches that decide what your language is without asking, and the gates that fail closed. Use when adding or extending language support, when touching…
running-tests
Cómo se ejecutan y se escriben los tests de Kivgraph - las tres suites, el tag ladybug que no se invoca a mano, qué se salta y por qué, y el smoke test del binario. Usar al correr tests, al añadirlos, ante un fallo de enlazado library 'lbug' not found, ante un SKIP inesperado, o antes de cerrar una tarea.
publishing-releases
Cómo y sobre todo cuándo se publica una release de Kivgraph - la versión sólo sube, los tres sitios donde vive, qué verifica CI, y la lista de cambios que NO merecen release. Usar al proponer o preparar una release, al tocar internal/version/version.go, al crear un tag vX.Y.Z, o cuando alguien pida "saca una versión".
kivgraph
Use Kivgraph's local MCP server to search symbols, read their source, and follow references, dependencies, cross-repository consumers and graph status in the indexed code graph.
opening-pull-requests
Open or update a Kivgraph pull request, wait for CI and CodeRabbit, verify review findings against the current code, apply valid feedback, and repeat until the latest head is review-ready. Use when the user asks to open, create, prepare, or finish a PR, or asks what CodeRabbit says about one.
wikipedia-glossary
Enrich an Open Mind project's glossary with standard term definitions from Wikipedia. For each already-extracted term, look up its authoritative Wikipedia definition, tidy it, and write it back to the glossary as a separate, attributed field; terms with no confident match keep their verbatim in-project definition. Use…