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 skills add AdityaVasireddy/agent-skills --skill engineering-historiangit clone --depth 1 https://github.com/AdityaVasireddy/agent-skillsWrote 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/adityavasireddy/agent-skills/engineering-historian)<a href="https://agentmods.dev/skills/adityavasireddy/agent-skills/engineering-historian"><img src="https://agentmods.dev/badge/skills/adityavasireddy/agent-skills/engineering-historian/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/adityavasireddy/agent-skills/engineering-historian"><img src="https://agentmods.dev/badge/skills/adityavasireddy/agent-skills/engineering-historian.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00187 | $0.04780 |
| Opus 5.5 | $0.00075 | $0.01912 |
| Sonnet 5.5 | $0.00037 | $0.00956 |
| Haiku 4.5 | $0.00019 | $0.00478 |
Grade A, and why
engineering-historian 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 1mo 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 — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Engineering Historian (v3.1)
Storage is stupid; retrieval is smart — meaning the storage layer stays deliberately dumb (plain markdown, minimal taxonomy) so that intelligence is applied at read and distillation time, not speculated into folder structures. And wisdom is earned through cited evidence, not asserted. This system captures specific decisions as CASES, distills their reasoning into JUDGMENTS, promotes recurring patterns into PRINCIPLES, and hardens checkable principles into STANDARDS. Nothing enters as a rule without a cited case behind it. A principle without provenance is invalid in this system: delete it on sight.
What v3 changes from v2.5: nothing in the pipeline, formats, promotion rules, retrieval, or vault layout. v3 only changes when capture fires. In v2.5 a human had to remember to type /history or /case; that is where the system died in practice. v3 replaces the human trigger with an automatic post-session sweep that drafts cases (status: auto) from the transcript, reviewed in one batch at /distill. Manual /case and /history still work as overrides. Everything below the "Automation layer" heading is the only new material; everything above it is v2.5 verbatim.
What v3.1 changes from v3: nothing in what is captured or how — only the sweep's pre-filter. v3.0 skipped sessions with few user turns, which silently dropped prompt-light agentic sessions (three prompts, hours of autonomous work). v3.1 replaces that significance guess with a triviality gate: skip only when every available deterministic signal — transcript size, user turns, git activity — proves the session trivial; otherwise sweep. When uncertain, capture: a false run costs one cheap NOTHING call, a false skip loses history. Rationale and rejected alternatives: docs/adr/ADR-001-capture-gate-triviality.md; release notes: CHANGELOG.md.
Three modes: capture (Tier 1, now sweep-driven), distillation (Tier 2, periodic), retrieval.
The pipeline
What ships with it
13 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.
- assets/capture-rules.seed.md 684 B
- assets/SWEEP.md 6.9 KB
- CHANGELOG.md 15 KB
- docs/adr/ADR-001-capture-gate-triviality.md 6.0 KB
- example/2026-07-05.md 7.9 KB
- example/retrieval-example.md 3.1 KB
- INSTALL.md 13 KB
- README.md 5.0 KB
- references/promotion-rules.md 7.4 KB
- references/templates.md 6.9 KB
- references/worked-examples.md 11 KB
- scripts/historian-sweep.sh 31 KB runs code
- tests/run-tests.sh 23 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.
- 1mo ago First seen · 179 lines · 187 tokens per session scan A 841c46f0ad8d
engineering-historian is a skill published in the GitHub repository AdityaVasireddy/agent-skills (2 stars, last pushed 1mo ago), licensed MIT. It adds 187 tokens to every session and 4,780 once invoked, about $0.0007 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
context-engine
Loads and manages company context for all C-suite advisor skills. Reads /.claude/company-context.md, detects stale context (older than 90 days), enriches context during conversations, and enforces privacy/anonymization rules before external API calls. Use when starting any C-suite advisor session, when context looks…
context-bankruptcy
Declare bankruptcy on a long-lived AI agent's accumulated memory — audit what it currently believes, separate ground truth from stale and wrong, purge deliberately, restate the truths that survive, and log what was lost. Use when an agent keeps acting on outdated facts, contradicts itself across sessions, 'remembers'…
context-engineering-review
Review what an LLM feature or agent actually puts in its context window — and find what's bloating, missing, or fighting itself. Use when asked to review a system prompt and context assembly, cut token usage without losing quality, debug an agent that ignores instructions, or audit how retrieval results, history, and…
memory-file-maintenance
Keep your AI memory/context file (MEMORY.md, CLAUDE.md, custom instructions) healthy over time — pruning the stale, adding the new, and keeping it sharp so your AI keeps getting you right. Use when asked review my memory file, my AI context is outdated, clean up my CLAUDE.md, or maintain my AI instructions. Produces a…
agent-project-file
Set up the project file an AI agent keeps for work that runs over days or weeks: the goal, the definition of done, milestones, the current state, decisions made, open questions, and the exact point to resume from after an interruption. Use when asked to plan a long-running agent task, make my agent work on this over a…
handoff
Create a structured session handoff when context is running low or work is pausing. Deep context mining, self-validation, multi-file splitting. Captures everything the next session needs.