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 commands/synaptic-labs-ai/pact-plugin/pin-memorygit clone --depth 1 https://github.com/Synaptic-Labs-AI/PACT-PluginWrote 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/commands/synaptic-labs-ai/pact-plugin/pin-memory)<a href="https://agentmods.dev/commands/synaptic-labs-ai/pact-plugin/pin-memory"><img src="https://agentmods.dev/badge/commands/synaptic-labs-ai/pact-plugin/pin-memory.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 | $0.00018 | $0.01439 |
| Opus 5 | $0.00009 | $0.00720 |
| Sonnet 5 | $0.00004 | $0.00288 |
| Haiku 4.5 | $0.00002 | $0.00144 |
Grade A, and why
pin-memory 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 — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Mode
- With arguments (
/PACT:pin-memory <content>): Pin the specified content. - Without arguments (
/PACT:pin-memory): Review the session for pin-worthy context and pin what matters.
Caps (enforced mechanically)
Cap violations are denied by hooks/pin_caps_gate.py when the Edit/Write tool call lands. You do NOT need to invoke a CLI check before adding — the hook is authoritative.
- Count: 12 pins maximum.
- Size: 1500 characters per pin body (excludes
<!-- pinned: ... -->and<!-- STALE: ... -->auto-markers). - Override: verbatim load-bearing content MAY carry a
pin-size-overriderationale (≤ 120 chars, single line) — see Size Override. The hook validates the rationale in-band.
When to Pin
- Critical gotchas that would waste hours if forgotten
- Key architectural decisions that explain "why" (not "what")
- Build/deploy commands needed every session
- Non-obvious patterns unique to this codebase
When NOT to Pin
- Routine session context (auto-memory and pact-memory handle this)
- Things easily found in code or docs
- Temporary information that will become stale
Process
Target file: The project CLAUDE.md may be at either $CLAUDE_PROJECT_DIR/.claude/CLAUDE.md (preferred) or $CLAUDE_PROJECT_DIR/CLAUDE.md (legacy). Use .claude/CLAUDE.md if it exists, otherwise ./CLAUDE.md. If neither exists, create at .claude/CLAUDE.md.
Adding a pin
- Read existing CLAUDE.md.
- Locate or create a
## Pinned Contextsection (place it before## Working Memory). - If the file carries a
<!-- PACT_MEMORY_PINNED_END -->line, the new entry MUST go ABOVE that line. The pinned region ends there. A pin placed below it sits outside the region, so the count and size caps do not measure it and the hook cannot deny it — the pin is silently uncapped. Insert immediately before that line, after the last existing entry. If the file has no such line, append at the end of the section as usual. - Add the new entry with a date tag. Make this write with the
Edittool. Do NOT use theWritetool. Step 1 gave you the full file, so aWritecall looks like the short route to the new content. After aWrite, a file that had CRLF line endings has LF line endings. The curator then sees a change to a file they did not edit. AnEditkeeps the line endings of the file. Use the example below:<!-- pinned: YYYY-MM-DD --> ### Entry Title Content here (~5-10 lines max) - Commit — then confirm
CLAUDE.mdis actually in the resulting commit. Where a project git-ignoresCLAUDE.md(a bareCLAUDE.mdline in.gitignorematches at any depth), the failure is loud in the obvious cases and silent in the one that matters. An explicitgit add CLAUDE.mderrors, and a commit whose only change isCLAUDE.mdreports "nothing to commit" — both non-zero, both visible. But a blanketgit commit -a/-amthat also picks up other changed files succeeds, exits 0, and simply omitsCLAUDE.md. Checking the commit's exit status instead of its contents is what turns that into a pin reported as durable while it has no git backing at all. So verify the file is in the commit, not that the commit succeeded. If it is not, report the pin as applied to the working file but not version-controlled, so the curator knows it lives only on this machine. NEVER usegit add -fto force it past the ignore rule — the rule is the project's decision, not an obstacle.
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 · 77 lines · 18 tokens per session scan A b7d787a0a85c
pin-memory is a command published in the GitHub repository Synaptic-Labs-AI/PACT-Plugin (71 stars, last pushed 3d ago), licensed MIT. It adds 18 tokens to every session and 1,439 once invoked, about $0.0001 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 commands, from other repositories
git
Git operations with intelligent commit messages and workflow optimization.
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.